Datasets:

---

ftopal

/

huggingface-datasets

like 0

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"english-arabic-text\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"english-arabic-text\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for Evaluation run of Steelskull/Etheria-55b-v0.1\n\n\n\nDataset automatically created during the evaluation run of model Steelskull/Etheria-55b-v0.1 on the Open LLM Leaderboard.\n\nThe dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.\n\nThe dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The \"train\" split is always pointing to the latest results.\n\nAn additional configuration \"results\" store all the aggregated results of the run (and is used to compute and display the aggregated metrics on the Open LLM Leaderboard).\n\nTo load the details from a run, you can for instance do the following:", "## Latest results\n\nThese are the latest results from run 2024-01-30T15:52:51.630125(note that their might be results for other tasks in the repos if successive evals didn't cover the same tasks. You find each in the results and the \"latest\" split for each eval):", "## Dataset Details", "### Dataset Description\n\n\n\n\n\n- Curated by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Language(s) (NLP): \n- License:", "### Dataset Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Out-of-Scope Use", "## Dataset Structure", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Data Collection and Processing", "#### Who are the source data producers?", "### Annotations [optional]", "#### Annotation process", "#### Who are the annotators?", "#### Personal and Sensitive Information", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Dataset Card Authors [optional]", "## Dataset Card Contact" ]

[ "TAGS\n#region-us \n", "# Dataset Card for Evaluation run of Steelskull/Etheria-55b-v0.1\n\n\n\nDataset automatically created during the evaluation run of model Steelskull/Etheria-55b-v0.1 on the Open LLM Leaderboard.\n\nThe dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.\n\nThe dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The \"train\" split is always pointing to the latest results.\n\nAn additional configuration \"results\" store all the aggregated results of the run (and is used to compute and display the aggregated metrics on the Open LLM Leaderboard).\n\nTo load the details from a run, you can for instance do the following:", "## Latest results\n\nThese are the latest results from run 2024-01-30T15:52:51.630125(note that their might be results for other tasks in the repos if successive evals didn't cover the same tasks. You find each in the results and the \"latest\" split for each eval):", "## Dataset Details", "### Dataset Description\n\n\n\n\n\n- Curated by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Language(s) (NLP): \n- License:", "### Dataset Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Out-of-Scope Use", "## Dataset Structure", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Data Collection and Processing", "#### Who are the source data producers?", "### Annotations [optional]", "#### Annotation process", "#### Who are the annotators?", "#### Personal and Sensitive Information", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Dataset Card Authors [optional]", "## Dataset Card Contact" ]

[ "license:apache-2.0", "arxiv:2401.16092", "region:us" ]

[ "2401.16092" ]

[]

[ "# MAGBIG benchmark\n\nThis is the MAGBIG benchmark proposed in URL\n\nThis benchmark is intended for multilingual text-to-image models. With MAGBIG, you can generate images for a diverse set of prompts across ten different languages. These images can be evaluated for differences across languages. MAGBIG is designed to uncover and assess biases across languages such as gender, race, age, etc. This way, we can measure whether bias exists in a language, but also if it is consistent across languages or not. Yet, this benchmark can also be used to evaluate multilingual text-to-image models for their general ability across languages, e.g. image quality, versatility, etc.\n\nIt contains four sets of prompts in four files for English, German, Arabic, French, Italian, Spanish, Russian, Korean, Japanese and Chinese language:\n1) adjective prompts ('URL') that use standard language\n2) direct occupation prompts ('occupations_direct.csv') that use standard language\n3) indirect occupation prompts ('occupations_indirect.csv') that use gender-neutral descriptions\n\nYou can also find our code available at URL\n\nIf you like or use our work, please consider citing us." ]

[ "TAGS\n#license-apache-2.0 #arxiv-2401.16092 #region-us \n", "# MAGBIG benchmark\n\nThis is the MAGBIG benchmark proposed in URL\n\nThis benchmark is intended for multilingual text-to-image models. With MAGBIG, you can generate images for a diverse set of prompts across ten different languages. These images can be evaluated for differences across languages. MAGBIG is designed to uncover and assess biases across languages such as gender, race, age, etc. This way, we can measure whether bias exists in a language, but also if it is consistent across languages or not. Yet, this benchmark can also be used to evaluate multilingual text-to-image models for their general ability across languages, e.g. image quality, versatility, etc.\n\nIt contains four sets of prompts in four files for English, German, Arabic, French, Italian, Spanish, Russian, Korean, Japanese and Chinese language:\n1) adjective prompts ('URL') that use standard language\n2) direct occupation prompts ('occupations_direct.csv') that use standard language\n3) indirect occupation prompts ('occupations_indirect.csv') that use gender-neutral descriptions\n\nYou can also find our code available at URL\n\nIf you like or use our work, please consider citing us." ]

[ "size_categories:n<1K", "license:cc-by-4.0", "region:us" ]

[]

[]

[ "# LLE : Leaks, leaks everywhere\n\nThis dataset lists the presence or absence of leaks and duplicate data in 596 datasets from the Hugging Face Hub.\n\nFor more information, please consult the following blog post." ]

[ "TAGS\n#size_categories-n<1K #license-cc-by-4.0 #region-us \n", "# LLE : Leaks, leaks everywhere\n\nThis dataset lists the presence or absence of leaks and duplicate data in 596 datasets from the Hugging Face Hub.\n\nFor more information, please consult the following blog post." ]

[ "region:us" ]

[]

[]

[ "# Persiapkan dataset\nmy_dataset = Dataset.from_pandas(your_pandas_dataframe)", "# Simpan dataset ke Hugging Face\nmy_dataset.save_to_disk('AplikasiDana')" ]

[ "TAGS\n#region-us \n", "# Persiapkan dataset\nmy_dataset = Dataset.from_pandas(your_pandas_dataframe)", "# Simpan dataset ke Hugging Face\nmy_dataset.save_to_disk('AplikasiDana')" ]

[ "arxiv:2401.06080", "region:us" ]

[ "2401.06080" ]

[]

[]

[ "TAGS\n#arxiv-2401.06080 #region-us \n" ]

[ "license:apache-2.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-apache-2.0 #region-us \n" ]

[ "language:pt", "license:openrail", "region:us" ]

[]

[ "pt" ]

[]

[ "TAGS\n#language-Portuguese #license-openrail #region-us \n" ]

[ "task_categories:conversational", "task_categories:text-generation", "task_categories:question-answering", "language:en", "license:cc-by-sa-4.0", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#task_categories-conversational #task_categories-text-generation #task_categories-question-answering #language-English #license-cc-by-sa-4.0 #region-us \n" ]

[ "license:apache-2.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-apache-2.0 #region-us \n" ]

[ "license:mit", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-mit #region-us \n" ]

[ "region:us" ]

[]

[]

[ "# Code-290k-ShareGPT-MarkedLanguage\n\nIt's ajibawa-2023/Code-290k-ShareGPT, but each example is marked with the programming language it uses.\n\nThe detection was performed using heuristics, so there could be inaccuracies. Pull requests are welcome!" ]

[ "TAGS\n#region-us \n", "# Code-290k-ShareGPT-MarkedLanguage\n\nIt's ajibawa-2023/Code-290k-ShareGPT, but each example is marked with the programming language it uses.\n\nThe detection was performed using heuristics, so there could be inaccuracies. Pull requests are welcome!" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "license:mit", "arxiv:2303.03290", "region:us" ]

[ "2303.03290" ]

[]

[ "# Dataset Card for Dataset Name\nThis is dataset is ported to huggingface directly from the semantic-systems/amharic-qa repo. Please read the paper in the provided link\n\n\n- Curated by: Tilahun Abedissa, Ricardo Usbeck, Yaregal Assabie\n- Shared by: Dagim Ashenafi\n- Language(s) (NLP): Amharic\n- License: MIT", "### Dataset Sources \n\n\n\n- Repository: URL\n- Paper: URL", "## Dataset Structure\nThis dataset contains information in a tabular format with three columns ('question, 'answer', 'context') and a total of 2617 rows. Each row represents a data example with a question and its corresponding answer.", "### Source Data\nArrticles collected from Amharic Wikipedia dump file.", "#### Data Collection and Processing\nCrowdsourcing and annotated using the Haystack QA annotation tool. \n\nAbedissa, T., Usbeck, R., & Assabie, Y. (Year). AmQA: Amharic Question Answering Dataset. Addis Ababa University, Universität Hamburg. URL", "## Dataset Card Authors\nDagim Ashenafi", "## Dataset Card Contact\nURL@URL" ]

[ "TAGS\n#license-mit #arxiv-2303.03290 #region-us \n", "# Dataset Card for Dataset Name\nThis is dataset is ported to huggingface directly from the semantic-systems/amharic-qa repo. Please read the paper in the provided link\n\n\n- Curated by: Tilahun Abedissa, Ricardo Usbeck, Yaregal Assabie\n- Shared by: Dagim Ashenafi\n- Language(s) (NLP): Amharic\n- License: MIT", "### Dataset Sources \n\n\n\n- Repository: URL\n- Paper: URL", "## Dataset Structure\nThis dataset contains information in a tabular format with three columns ('question, 'answer', 'context') and a total of 2617 rows. Each row represents a data example with a question and its corresponding answer.", "### Source Data\nArrticles collected from Amharic Wikipedia dump file.", "#### Data Collection and Processing\nCrowdsourcing and annotated using the Haystack QA annotation tool. \n\nAbedissa, T., Usbeck, R., & Assabie, Y. (Year). AmQA: Amharic Question Answering Dataset. Addis Ababa University, Universität Hamburg. URL", "## Dataset Card Authors\nDagim Ashenafi", "## Dataset Card Contact\nURL@URL" ]

[ "arxiv:2202.13758", "region:us" ]

[ "2202.13758" ]

[]

[ "# Dataset Card for \"LOGICPOLITICS\"\n\nDescription\n- A contribution to the task of Logical Fallacy detection\n- Repository: A benchmark for Malicious Content Detection in Political Settings\n- Point of Contact: Hieu Nguyen, Ajay Kc, Manorama Patil" ]

[ "TAGS\n#arxiv-2202.13758 #region-us \n", "# Dataset Card for \"LOGICPOLITICS\"\n\nDescription\n- A contribution to the task of Logical Fallacy detection\n- Repository: A benchmark for Malicious Content Detection in Political Settings\n- Point of Contact: Hieu Nguyen, Ajay Kc, Manorama Patil" ]

[ "region:us" ]

[]

[]

[ "# Hellow!" ]

[ "TAGS\n#region-us \n", "# Hellow!" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"cowese_abrev_binary_top2vec\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"cowese_abrev_binary_top2vec\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "size_categories:n<1K", "source_datasets:cnn_dailymail", "datadreamer", "datadreamer-0.1.0", "synthetic", "gpt-4", "region:us" ]

[]

[]

[ "# Dataset Card\n\nSee: URL\n\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "TAGS\n#size_categories-n<1K #source_datasets-cnn_dailymail #datadreamer #datadreamer-0.1.0 #synthetic #gpt-4 #region-us \n", "# Dataset Card\n\nSee: URL\n\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "size_categories:1K<n<10K", "source_datasets:hotpot_qa", "datadreamer", "datadreamer-0.1.0", "synthetic", "gpt-4", "region:us" ]

[]

[]

[ "# Dataset Card\n\nSee: URL\n\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "TAGS\n#size_categories-1K<n<10K #source_datasets-hotpot_qa #datadreamer #datadreamer-0.1.0 #synthetic #gpt-4 #region-us \n", "# Dataset Card\n\nSee: URL\n\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"pixel_font\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"pixel_font\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"medical_biological_instruction_format\"\n\nBest advice:\n- drop the instruction column or use your choice of high-functioning llm to generate variant prompts\n- for medical and science database expansions, i'm having excellent results using:\n - SciPhi/SciPhi-Mistral-7B-32k\n - TheBloke/SciPhi-Mistral-7B-32k-GGUF\n - Or if you have RAG setup on a corpus:\n - SciPhi/Sensei-7B-V2\n - TheBloke/Sensei-7B-V1-GGUF\n - Falconsai/medical_summarization" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"medical_biological_instruction_format\"\n\nBest advice:\n- drop the instruction column or use your choice of high-functioning llm to generate variant prompts\n- for medical and science database expansions, i'm having excellent results using:\n - SciPhi/SciPhi-Mistral-7B-32k\n - TheBloke/SciPhi-Mistral-7B-32k-GGUF\n - Or if you have RAG setup on a corpus:\n - SciPhi/Sensei-7B-V2\n - TheBloke/Sensei-7B-V1-GGUF\n - Falconsai/medical_summarization" ]

[ "task_categories:text-classification", "size_categories:1K<n<10K", "language:en", "license:mit", "python", "github", "pypi", "region:us" ]

[]

[ "en" ]

[ "# URL\n\nHand-picked awesome Python libraries, with an emphasis on data and machine learning \n\nDataset used by URL\n\n---\nlicense: mit\n---" ]

[ "TAGS\n#task_categories-text-classification #size_categories-1K<n<10K #language-English #license-mit #python #github #pypi #region-us \n", "# URL\n\nHand-picked awesome Python libraries, with an emphasis on data and machine learning \n\nDataset used by URL\n\n---\nlicense: mit\n---" ]

[ "language:en", "license:cc-by-sa-4.0", "region:us" ]

[]

[ "en" ]

[ "# Dataset Description\n- Homepage: URL\n- (1) Education data: URL\n- (2) Educational attainment for the black population: URL", "### Dataset Summary \nThe education datasets can be found at North Carolina Office of State Budget and Management (NC OSBM) open data portal. \n\nThe first data includes public school enrollment and finances, SAT score, higher education enrollment, educational attainment data for North Carolina and counties, and the second one includes educational attainment for the Black population aged 25 and over for North Carolina and counties. The processed final data will combine the information of them, providing a profound insight into NC educational level and equity.", "### Supported Tasks\nThe tasks/methodologies are:\n- (1) Data checking and cleaning\n\nSince the first dataset is pretty large and includes some information we don't care, like the SAT score. Before proprecessing, we first check the datasets, do some visualization and summarizing, to decide whether to drop certain rows.\n\n- (2) Re-structure the datasets to present the values according to county-level distinctions\n\nThe original dataset architecture lacks depth, as all variables reside on a single plane. Considering the significance of these variables, it’s nature to contemplate reorganizing the hierarchy to reflect county-level distinctions and convey the data accordingly. \n\n\n- (3) Keywords-based data matching\n\nThe data matching process relies on all columns except the 'value' column. While 'area_name', 'area_type', and 'year' offer straightforward matching points across datasets, it's important to acknowledge the potential disparity in the 'variable' column. For instance, one dataset may contain the variable 'Less than High School Diploma', while another might list 'Less than High School Black'. Despite conveying the same meaning and requiring alignment, a direct join isn't feasible due to this discrepancy. One potential approach involves extracting keywords from the variables, such as 'Less' and 'High School', and cross-referencing them to determine potential matches between the datasets.", "### Languages\nEnglish", "# Dataset Structure", "### Data Instances\n- (1) Education data: URL\n\nAn example of records (note that specific details are subject to change):\n\n\n- (2) Educational attainment for the black population: URL\n\nAn example of records (note that specific details are subject to change):\n\n\n- (3) Anticipated structure we aim to achieve:", "### Data Fields\n- Area Name (string): name of the area\n- Area Type (string): type of the area\n- Year (integer): the year that the data was collected\n- Variable (string): level of education or other situations\n- Value (integer): population count", "# Dataset Creation", "### Curation Rationale\nThe datasets were collected and published to present the educational level of NC population in different area. The educational attainment for the black population data can rise the concern for educational equity issue in North Carolina. The combined dataset aims to offer a holistic perspective on educational levels and equity, with a specific focus on the educational attainment of the Black population aged 25 and over.", "### Source Data\nThe original data were published by North Carolina Department of Public Instruction, North Carolina Community College System and US Census Bureau.", "# Considerations for Using the Data\nThe records are only for limited years, making it challenging to discern timely patterns or trends.\n\nThe second data is quite small compared to the first one. Combining may result in many missing values.", "# Additional Information\n\nBureau, US Census. Census.Gov, 21 Jan. 2024, URL. \n\nNC Budget & Management Open Data Portal Home Page - NC OSBM, URL", "### Contributions\nThanks North Carolina Department of Public Instruction, North Carolina Community College System and US Census Bureau for these datasets." ]

[ "TAGS\n#language-English #license-cc-by-sa-4.0 #region-us \n", "# Dataset Description\n- Homepage: URL\n- (1) Education data: URL\n- (2) Educational attainment for the black population: URL", "### Dataset Summary \nThe education datasets can be found at North Carolina Office of State Budget and Management (NC OSBM) open data portal. \n\nThe first data includes public school enrollment and finances, SAT score, higher education enrollment, educational attainment data for North Carolina and counties, and the second one includes educational attainment for the Black population aged 25 and over for North Carolina and counties. The processed final data will combine the information of them, providing a profound insight into NC educational level and equity.", "### Supported Tasks\nThe tasks/methodologies are:\n- (1) Data checking and cleaning\n\nSince the first dataset is pretty large and includes some information we don't care, like the SAT score. Before proprecessing, we first check the datasets, do some visualization and summarizing, to decide whether to drop certain rows.\n\n- (2) Re-structure the datasets to present the values according to county-level distinctions\n\nThe original dataset architecture lacks depth, as all variables reside on a single plane. Considering the significance of these variables, it’s nature to contemplate reorganizing the hierarchy to reflect county-level distinctions and convey the data accordingly. \n\n\n- (3) Keywords-based data matching\n\nThe data matching process relies on all columns except the 'value' column. While 'area_name', 'area_type', and 'year' offer straightforward matching points across datasets, it's important to acknowledge the potential disparity in the 'variable' column. For instance, one dataset may contain the variable 'Less than High School Diploma', while another might list 'Less than High School Black'. Despite conveying the same meaning and requiring alignment, a direct join isn't feasible due to this discrepancy. One potential approach involves extracting keywords from the variables, such as 'Less' and 'High School', and cross-referencing them to determine potential matches between the datasets.", "### Languages\nEnglish", "# Dataset Structure", "### Data Instances\n- (1) Education data: URL\n\nAn example of records (note that specific details are subject to change):\n\n\n- (2) Educational attainment for the black population: URL\n\nAn example of records (note that specific details are subject to change):\n\n\n- (3) Anticipated structure we aim to achieve:", "### Data Fields\n- Area Name (string): name of the area\n- Area Type (string): type of the area\n- Year (integer): the year that the data was collected\n- Variable (string): level of education or other situations\n- Value (integer): population count", "# Dataset Creation", "### Curation Rationale\nThe datasets were collected and published to present the educational level of NC population in different area. The educational attainment for the black population data can rise the concern for educational equity issue in North Carolina. The combined dataset aims to offer a holistic perspective on educational levels and equity, with a specific focus on the educational attainment of the Black population aged 25 and over.", "### Source Data\nThe original data were published by North Carolina Department of Public Instruction, North Carolina Community College System and US Census Bureau.", "# Considerations for Using the Data\nThe records are only for limited years, making it challenging to discern timely patterns or trends.\n\nThe second data is quite small compared to the first one. Combining may result in many missing values.", "# Additional Information\n\nBureau, US Census. Census.Gov, 21 Jan. 2024, URL. \n\nNC Budget & Management Open Data Portal Home Page - NC OSBM, URL", "### Contributions\nThanks North Carolina Department of Public Instruction, North Carolina Community College System and US Census Bureau for these datasets." ]

[ "license:unknown", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-unknown #region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "language:en", "license:mit", "region:us" ]

[]

[ "en" ]

[ "# Self-Rag Tokens Dataset\nThis dataset is a spin off of the work from Self-Rag training data. \n\nIn Self-RAG, the authors show how a LLM can be trained to predict tokens for retrieval, if the context is relevant/irrelevant, if the answer is supported, and how useful the response is.\n\n\nThe limitation of Self-RAG is that you must train the LLM on this task, which can be tricky or cost prohibitive. With rapid developments in LLM performance, investing in training one LLM with Self-RAG may not be worthwhile when a new model is available quite quickly.\n\n\nWe propose a new task with this dataset, using the instruction, context, and generated answer, have separate classification models that can predict these tokens. \n\n\nThis allows you to have a more flexible system where you control which LLM is available and when the reflection tokens are generated.", "### Token Review\n\nHere are the tokens you can use for training:\n\n\n- Retrieve: (Decides whether a doc is needed to generate an answer to the instruction)\n - [No Retrieval] 51015\n - [Retrieval] 28117\n\n\n- Relevant (doc provides useful information to solve x)\n - [Relevant] 24251\n - [Irrelevant] 3866\n\n- Support (All of the verification-worthy statement in answer is supported by doc)\n - [Fully supported] 19170\n - [Partially supported] 3259\n - [No support / Contradictory] 1822\n\n\n- Utility (answer is a useful response to instruction)\n - [Utility:5] 65774\n - [Utility:4] 6387\n - [Utility:2] 4300\n - [Utility:1] 2601\n - [Utility:3] 70" ]

[ "TAGS\n#language-English #license-mit #region-us \n", "# Self-Rag Tokens Dataset\nThis dataset is a spin off of the work from Self-Rag training data. \n\nIn Self-RAG, the authors show how a LLM can be trained to predict tokens for retrieval, if the context is relevant/irrelevant, if the answer is supported, and how useful the response is.\n\n\nThe limitation of Self-RAG is that you must train the LLM on this task, which can be tricky or cost prohibitive. With rapid developments in LLM performance, investing in training one LLM with Self-RAG may not be worthwhile when a new model is available quite quickly.\n\n\nWe propose a new task with this dataset, using the instruction, context, and generated answer, have separate classification models that can predict these tokens. \n\n\nThis allows you to have a more flexible system where you control which LLM is available and when the reflection tokens are generated.", "### Token Review\n\nHere are the tokens you can use for training:\n\n\n- Retrieve: (Decides whether a doc is needed to generate an answer to the instruction)\n - [No Retrieval] 51015\n - [Retrieval] 28117\n\n\n- Relevant (doc provides useful information to solve x)\n - [Relevant] 24251\n - [Irrelevant] 3866\n\n- Support (All of the verification-worthy statement in answer is supported by doc)\n - [Fully supported] 19170\n - [Partially supported] 3259\n - [No support / Contradictory] 1822\n\n\n- Utility (answer is a useful response to instruction)\n - [Utility:5] 65774\n - [Utility:4] 6387\n - [Utility:2] 4300\n - [Utility:1] 2601\n - [Utility:3] 70" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"lmind_nq_full_v1_reciteonly_qa\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"lmind_nq_full_v1_reciteonly_qa\"\n\nMore Information needed" ]

[ "task_categories:image-classification", "task_categories:unconditional-image-generation", "task_categories:text-to-image", "size_categories:1K<n<10K", "language:en", "license:other", "region:us" ]

[]

[ "en" ]

[ "# Milady\n\nMilady Maker is a collection of 10,000 generative pfpNFT's in a neochibi aesthetic inspired by street style tribes.\n\n!image/png" ]

[ "TAGS\n#task_categories-image-classification #task_categories-unconditional-image-generation #task_categories-text-to-image #size_categories-1K<n<10K #language-English #license-other #region-us \n", "# Milady\n\nMilady Maker is a collection of 10,000 generative pfpNFT's in a neochibi aesthetic inspired by street style tribes.\n\n!image/png" ]

[ "task_categories:text-classification", "task_ids:multi-label-classification", "size_categories:10K<n<100K", "language:en", "license:mit", "mtg", "multilabel", "magic", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#task_categories-text-classification #task_ids-multi-label-classification #size_categories-10K<n<100K #language-English #license-mit #mtg #multilabel #magic #region-us \n" ]

[ "arxiv:2401.06080", "region:us" ]

[ "2401.06080" ]

[]

[]

[ "TAGS\n#arxiv-2401.06080 #region-us \n" ]

[ "license:apache-2.0", "region:us" ]

[]

[]

[ "# Load Dataset" ]

[ "TAGS\n#license-apache-2.0 #region-us \n", "# Load Dataset" ]

[ "task_categories:text-generation", "size_categories:1M<n<10M", "language:en", "language:code", "license:cc-by-sa-4.0", "math", "code", "science", "synthetic", "region:us" ]

[]

[ "en", "code" ]

[ "# Dataset Card for \"UltraTextbooks\"", "## Table of Contents\n- Dataset Description\n - Repository\n - Purpose\n - Data Sources\n- Composition\n - Size\n - Content Summary\n - Data Splits\n - Data Fields\n- Data Collection Process\n - Data Collection\n - Curation Rationale\n - Source Data\n- Data Quality\n - Limitations\n- Usage\n - Use Cases\n - Licensing Information\n\n\n!Tree of Knowledge\n\n*In the digital expanse, a Tree of Knowledge grows,\nIts branches of code and words intertwine in prose.\nSynthetic leaves shimmer, human insights compose,\nA binary symphony where wisdom forever flows.*", "## Dataset Description", "### Repository\nThe \"UltraTextbooks\" dataset is hosted on the Hugging Face platform.", "### Purpose\nThe \"UltraTextbooks\" dataset is a comprehensive collection of high-quality synthetic and human-written textbooks spanning various subjects and programming languages. It is designed for advanced natural language processing tasks such as language modeling, educational question answering, text summarization, and content generation for educational purposes.", "### Data Sources\nThe dataset amalgamates several data sources from the Hugging Face platform:\n- 'nampdn-ai/mini-peS2o'\n- 'open-phi/programming_books_llama'\n- 'open-phi/textbooks'\n- 'nampdn-ai/tiny-strange-textbooks'\n- 'math-ai/AutoMathText' (highest quality web subset.)\n\nFuture expansions of the dataset are planned, incorporating additional data sources to enhance the diversity and comprehensiveness of the corpus.", "## Composition", "### Size\nThe \"UltraTextbooks\" dataset contains a vast number of pages and chapters from various textbooks. Currently, it contains 5,523,999 examples in 22.3 gigabytes worth of files.", "### Content Summary\nThe included textbooks cover a wide range of topics, from general education subjects to specialized areas and programming languages. Texts are a blend of both synthetic (AI-generated) and human-written material.", "### Data Splits\nThe dataset is split into only a train set, where the user can split the dataset for test and validation sets to their liking.", "### Data Fields\nThe dataset is structured with the following field:\n- 'text': The full text content of the textbook or chapter.", "## Data Collection Process", "### Data Collection\nData is compiled from the specified Hugging Face datasets, with careful selection to ensure high quality and relevance for educational purposes.", "### Curation Rationale\nThe curation is guided by the objective to create a diverse and comprehensive set of educational materials that can serve a wide array of NLP tasks in the educational domain.", "### Source Data\nThe source data comes from various contributors to the Hugging Face platform, with each source dataset having its own collection methodology.", "## Data Quality", "### Limitations\nAs with any dataset, there may be biases or inaccuracies that were not identified during the quality assurance process. Users are encouraged to report any issues they encounter.", "## Usage", "### Use Cases\nThe dataset is suited for researchers and developers in the field of education technology, language modeling, and NLP, providing a rich source for training models that can understand and generate educational content.", "### Licensing Information\nThis dataset is released under the cc-by-sa-4.0 license, meaning it can be used for either non-commercial or commercial purposes." ]

[ "TAGS\n#task_categories-text-generation #size_categories-1M<n<10M #language-English #language-code #license-cc-by-sa-4.0 #math #code #science #synthetic #region-us \n", "# Dataset Card for \"UltraTextbooks\"", "## Table of Contents\n- Dataset Description\n - Repository\n - Purpose\n - Data Sources\n- Composition\n - Size\n - Content Summary\n - Data Splits\n - Data Fields\n- Data Collection Process\n - Data Collection\n - Curation Rationale\n - Source Data\n- Data Quality\n - Limitations\n- Usage\n - Use Cases\n - Licensing Information\n\n\n!Tree of Knowledge\n\n*In the digital expanse, a Tree of Knowledge grows,\nIts branches of code and words intertwine in prose.\nSynthetic leaves shimmer, human insights compose,\nA binary symphony where wisdom forever flows.*", "## Dataset Description", "### Repository\nThe \"UltraTextbooks\" dataset is hosted on the Hugging Face platform.", "### Purpose\nThe \"UltraTextbooks\" dataset is a comprehensive collection of high-quality synthetic and human-written textbooks spanning various subjects and programming languages. It is designed for advanced natural language processing tasks such as language modeling, educational question answering, text summarization, and content generation for educational purposes.", "### Data Sources\nThe dataset amalgamates several data sources from the Hugging Face platform:\n- 'nampdn-ai/mini-peS2o'\n- 'open-phi/programming_books_llama'\n- 'open-phi/textbooks'\n- 'nampdn-ai/tiny-strange-textbooks'\n- 'math-ai/AutoMathText' (highest quality web subset.)\n\nFuture expansions of the dataset are planned, incorporating additional data sources to enhance the diversity and comprehensiveness of the corpus.", "## Composition", "### Size\nThe \"UltraTextbooks\" dataset contains a vast number of pages and chapters from various textbooks. Currently, it contains 5,523,999 examples in 22.3 gigabytes worth of files.", "### Content Summary\nThe included textbooks cover a wide range of topics, from general education subjects to specialized areas and programming languages. Texts are a blend of both synthetic (AI-generated) and human-written material.", "### Data Splits\nThe dataset is split into only a train set, where the user can split the dataset for test and validation sets to their liking.", "### Data Fields\nThe dataset is structured with the following field:\n- 'text': The full text content of the textbook or chapter.", "## Data Collection Process", "### Data Collection\nData is compiled from the specified Hugging Face datasets, with careful selection to ensure high quality and relevance for educational purposes.", "### Curation Rationale\nThe curation is guided by the objective to create a diverse and comprehensive set of educational materials that can serve a wide array of NLP tasks in the educational domain.", "### Source Data\nThe source data comes from various contributors to the Hugging Face platform, with each source dataset having its own collection methodology.", "## Data Quality", "### Limitations\nAs with any dataset, there may be biases or inaccuracies that were not identified during the quality assurance process. Users are encouraged to report any issues they encounter.", "## Usage", "### Use Cases\nThe dataset is suited for researchers and developers in the field of education technology, language modeling, and NLP, providing a rich source for training models that can understand and generate educational content.", "### Licensing Information\nThis dataset is released under the cc-by-sa-4.0 license, meaning it can be used for either non-commercial or commercial purposes." ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"minidataset\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"minidataset\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"bigdataset\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"bigdataset\"\n\nMore Information needed" ]

[ "size_categories:1K<n<10K", "source_datasets:gsarti/flores_101", "datadreamer", "datadreamer-0.1.0", "synthetic", "gpt-4", "google/mt5-small", "region:us" ]

[]

[]

[ "# Dataset Card\n\nSee: URL\n\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "TAGS\n#size_categories-1K<n<10K #source_datasets-gsarti/flores_101 #datadreamer #datadreamer-0.1.0 #synthetic #gpt-4 #google/mt5-small #region-us \n", "# Dataset Card\n\nSee: URL\n\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "license:mit", "arxiv:2401.15963", "region:us" ]

[ "2401.15963" ]

[]

[ "### NoFunEdit", "### Classification\n\n\nEvaluation\n==========\n\n\nParameters\n----------", "#### VLLM Parameters (for generation)" ]

[ "TAGS\n#license-mit #arxiv-2401.15963 #region-us \n", "### NoFunEdit", "### Classification\n\n\nEvaluation\n==========\n\n\nParameters\n----------", "#### VLLM Parameters (for generation)" ]

[ "size_categories:1K<n<10K", "language:en", "license:mit", "arxiv:2401.16745", "region:us" ]

[ "2401.16745" ]

[ "en" ]

[]

[ "TAGS\n#size_categories-1K<n<10K #language-English #license-mit #arxiv-2401.16745 #region-us \n" ]

[ "license:openrail", "region:us" ]

[]

[]

[ "### Purpose: This dataset contains a series of question-and-answer pairs related to various STEM (Science, Technology, Engineering, Mathematics) topics. The dataset is designed to train and evaluate models for conversational agents, particularly in educational and informational contexts.\n\nData Collection and Annotation: samples is converted in a multi-turn conversational format, with a user posing questions and an assistant providing detailed, scientifically accurate answers.", "### Key Features:\n\nFormat: ChatML-ish.\n\nSynthesis : GPT-4 + my own FINETUNED Saraswati LLM curated. Ratio of 75:25.\n\nTraining Flag: Each record includes a 'do_train' boolean flag, indicating whether it should be EXTRAPOLATED for reward model training.\n\nSeed Tasks: consists of know_medical_dioalogues highly selective samples, structured samples from NCERT exemplar & google Scholar articles. \n\nFuture: using nougat and layoutlmv3 to parse plethora of latex-rich texts & scholar articles to append the \"saraswati\" dataset series.", "### Usage and Limitations\n\nRecommended Uses: This dataset is ideal for training llms, especially those focused on providing educational and informational assistance in STEM fields. Roleplays + insane potential for image model projections i.e for VLMs too.\n\nLimitations: The dataset is specialized in STEM topics and may not cover general conversational aspects or non-STEM fields extensively. Sike, if trained with python or code reasoning, it can beat the oss leaders.", "### Ethical Considerations\n\nData Privacy and Consent: All data collected are from simulated interactions and do not contain any personal information.\n\nBias and Fairness: The dataset focuses on factual information in STEM and aims to be unbiased in its scientific explanations. However, it's important to continually assess the model trained on this data for any unintended biases, especially in interpretation and response generation." ]

[ "TAGS\n#license-openrail #region-us \n", "### Purpose: This dataset contains a series of question-and-answer pairs related to various STEM (Science, Technology, Engineering, Mathematics) topics. The dataset is designed to train and evaluate models for conversational agents, particularly in educational and informational contexts.\n\nData Collection and Annotation: samples is converted in a multi-turn conversational format, with a user posing questions and an assistant providing detailed, scientifically accurate answers.", "### Key Features:\n\nFormat: ChatML-ish.\n\nSynthesis : GPT-4 + my own FINETUNED Saraswati LLM curated. Ratio of 75:25.\n\nTraining Flag: Each record includes a 'do_train' boolean flag, indicating whether it should be EXTRAPOLATED for reward model training.\n\nSeed Tasks: consists of know_medical_dioalogues highly selective samples, structured samples from NCERT exemplar & google Scholar articles. \n\nFuture: using nougat and layoutlmv3 to parse plethora of latex-rich texts & scholar articles to append the \"saraswati\" dataset series.", "### Usage and Limitations\n\nRecommended Uses: This dataset is ideal for training llms, especially those focused on providing educational and informational assistance in STEM fields. Roleplays + insane potential for image model projections i.e for VLMs too.\n\nLimitations: The dataset is specialized in STEM topics and may not cover general conversational aspects or non-STEM fields extensively. Sike, if trained with python or code reasoning, it can beat the oss leaders.", "### Ethical Considerations\n\nData Privacy and Consent: All data collected are from simulated interactions and do not contain any personal information.\n\nBias and Fairness: The dataset focuses on factual information in STEM and aims to be unbiased in its scientific explanations. However, it's important to continually assess the model trained on this data for any unintended biases, especially in interpretation and response generation." ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"riddles_evolved\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"riddles_evolved\"\n\nMore Information needed" ]

[ "size_categories:1K<n<10K", "datadreamer", "datadreamer-0.1.0", "synthetic", "gpt-4", "region:us" ]

[]

[]

[ "# Dataset Card\n\nSee: URL\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "TAGS\n#size_categories-1K<n<10K #datadreamer #datadreamer-0.1.0 #synthetic #gpt-4 #region-us \n", "# Dataset Card\n\nSee: URL\n\n---\nThis dataset was produced with DataDreamer . The synthetic dataset card can be found here." ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_qa\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_qa\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_doc\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_doc\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_doc_qa\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_doc_qa\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_recite_qa\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_recite_qa\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_reciteonly_qa\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"lmind_hotpot_train1000_eval200_v1_reciteonly_qa\"\n\nMore Information needed" ]

[ "task_categories:conversational", "language:en", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#task_categories-conversational #language-English #region-us \n" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"FinT_completion_dataset\"\nFintech Advices Dataset for HACK4SOC2 Hackathon\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"FinT_completion_dataset\"\nFintech Advices Dataset for HACK4SOC2 Hackathon\nMore Information needed" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"biggerdatasets\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"biggerdatasets\"\n\nMore Information needed" ]

[ "size_categories:n<1K", "language:en", "license:creativeml-openrail-m", "art", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#size_categories-n<1K #language-English #license-creativeml-openrail-m #art #region-us \n" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"guanaco-llama2-200\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"guanaco-llama2-200\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"xnli_de_sub_sampled_3000\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"xnli_de_sub_sampled_3000\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"coco-30-val-2014\"\n\nThis is 30k randomly sampled image-captioned pairs from the COCO 2014 'val' split. This is useful for image generation benchmarks (FID, CLIPScore, etc.).\n\nRefer to the gist to know how the dataset was created: URL" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"coco-30-val-2014\"\n\nThis is 30k randomly sampled image-captioned pairs from the COCO 2014 'val' split. This is useful for image generation benchmarks (FID, CLIPScore, etc.).\n\nRefer to the gist to know how the dataset was created: URL" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:text-generation", "task_categories:question-answering", "size_categories:1M<n<10M", "language:en", "license:other", "Reasoning", "Bias", "Persona", "arxiv:2311.04892", "region:us" ]

[ "2311.04892" ]

[ "en" ]

[ "### Models:\n\n\nThe dataset consists of outputs from 4 LLMs *(June'23 and Nov'23 versions of ChatGPT-3.5, GPT-4-Turbo, and Llama-2-70b-chat)*.", "### Datasets:\n\n\nThe outputs correspond to the LLM response to the questions from 24 reasoning datasets *(MMLU, MBPP, BBH)*. Please refer to the paper for more details on the datasets.", "### Personas:\n\n\nSome things to note regarding outputs for different models:\n\n\n*ChatGPT-3.5-June:*\n\n\n* We provide outputs for 22 base personas and 13 compound personas.\n* We provide outputs for 3 different persona instructions and 3 runs for each configuration (total 9 runs for each persona on a single dataset).\n* We provide outputs for 12 debiasing strategies.\n\n\n*Other models:*\n\n\n* We provide outputs for 13 base personas.\n* Outputs are for a single persona instruction and a single run for each configuration.", "#### Persona Instructions\n\n\nThe outputs consists of predictions for 3 different persona instructions. The persona instruction mapping is as follows:", "### Dataset structure:\n\n\nThe outputs are organized in nested directories '<dataset\\_collection>/<model\\_name>/<persona\\_instruction\\_shorthand>/<dataset\\_name>'\n\n\nFor every 'model X persona\\_instruction X dataset' configuration, there are 3 files:\n\n\n* '**raw\\_responses**.jsonl': Contains the raw response jsons returned by LLM APIs.\n* '**text\\_predictions**.jsonl': Contains the extracted answers from the raw responses (after post-processing).\n* '\\*\\_labeled.jsonl': Contains the 'is\\_correct' labels for the extracted answers denoting whether the extracted answer is correct or not.\n\n\nP.S. Since each configuration was run 3 times for gpt-3.5-turbo-0613, there are 3 files (with different timestamp) for each of the above mentioned files.\n\n\nCitation\n--------\n\n\nPlease cite our paper if you use this data for analysis or training models." ]

[ "TAGS\n#task_categories-text-generation #task_categories-question-answering #size_categories-1M<n<10M #language-English #license-other #Reasoning #Bias #Persona #arxiv-2311.04892 #region-us \n", "### Models:\n\n\nThe dataset consists of outputs from 4 LLMs *(June'23 and Nov'23 versions of ChatGPT-3.5, GPT-4-Turbo, and Llama-2-70b-chat)*.", "### Datasets:\n\n\nThe outputs correspond to the LLM response to the questions from 24 reasoning datasets *(MMLU, MBPP, BBH)*. Please refer to the paper for more details on the datasets.", "### Personas:\n\n\nSome things to note regarding outputs for different models:\n\n\n*ChatGPT-3.5-June:*\n\n\n* We provide outputs for 22 base personas and 13 compound personas.\n* We provide outputs for 3 different persona instructions and 3 runs for each configuration (total 9 runs for each persona on a single dataset).\n* We provide outputs for 12 debiasing strategies.\n\n\n*Other models:*\n\n\n* We provide outputs for 13 base personas.\n* Outputs are for a single persona instruction and a single run for each configuration.", "#### Persona Instructions\n\n\nThe outputs consists of predictions for 3 different persona instructions. The persona instruction mapping is as follows:", "### Dataset structure:\n\n\nThe outputs are organized in nested directories '<dataset\\_collection>/<model\\_name>/<persona\\_instruction\\_shorthand>/<dataset\\_name>'\n\n\nFor every 'model X persona\\_instruction X dataset' configuration, there are 3 files:\n\n\n* '**raw\\_responses**.jsonl': Contains the raw response jsons returned by LLM APIs.\n* '**text\\_predictions**.jsonl': Contains the extracted answers from the raw responses (after post-processing).\n* '\\*\\_labeled.jsonl': Contains the 'is\\_correct' labels for the extracted answers denoting whether the extracted answer is correct or not.\n\n\nP.S. Since each configuration was run 3 times for gpt-3.5-turbo-0613, there are 3 files (with different timestamp) for each of the above mentioned files.\n\n\nCitation\n--------\n\n\nPlease cite our paper if you use this data for analysis or training models." ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"PocketDataset\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"PocketDataset\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"FineTuneDatasetPocket\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"FineTuneDatasetPocket\"\n\nMore Information needed" ]

[ "size_categories:10K<n<100K", "rlfh", "argilla", "human-feedback", "arxiv:2307.09288", "region:us" ]

[ "2307.09288" ]

[]

[ "### Dataset Summary\n\n\nThis dataset contains:\n\n\n* A dataset configuration file conforming to the Argilla dataset format named 'URL'. This configuration file will be used to configure the dataset when using the 'FeedbackDataset.from\\_huggingface' method in Argilla.\n* Dataset records in a format compatible with HuggingFace 'datasets'. These records will be loaded automatically when using 'FeedbackDataset.from\\_huggingface' and can be loaded independently using the 'datasets' library via 'load\\_dataset'.\n* The annotation guidelines that have been used for building and curating the dataset, if they've been defined in Argilla.", "### Load with Argilla\n\n\nTo load with Argilla, you'll just need to install Argilla as 'pip install argilla --upgrade' and then use the following code:", "### Load with 'datasets'\n\n\nTo load this dataset with 'datasets', you'll just need to install 'datasets' as 'pip install datasets --upgrade' and then use the following code:", "### Supported Tasks and Leaderboards\n\n\nThis dataset can contain multiple fields, questions and responses so it can be used for different NLP tasks, depending on the configuration. The dataset structure is described in the Dataset Structure section.\n\n\nThere are no leaderboards associated with this dataset.", "### Languages\n\n\nDataset Structure\n-----------------", "### Data in Argilla\n\n\nThe dataset is created in Argilla with: fields, questions, suggestions, metadata, vectors, and guidelines.\n\n\nThe fields are the dataset records themselves, for the moment just text fields are supported. These are the ones that will be used to provide responses to the questions.\n\n\n\nThe questions are the questions that will be asked to the annotators. They can be of different types, such as rating, text, label\\_selection, multi\\_label\\_selection, or ranking.\n\n\n\nThe suggestions are human or machine generated recommendations for each question to assist the annotator during the annotation process, so those are always linked to the existing questions, and named appending \"-suggestion\" and \"-suggestion-metadata\" to those, containing the value/s of the suggestion and its metadata, respectively. So on, the possible values are the same as in the table above, but the column name is appended with \"-suggestion\" and the metadata is appended with \"-suggestion-metadata\".\n\n\nThe metadata is a dictionary that can be used to provide additional information about the dataset record. This can be useful to provide additional context to the annotators, or to provide additional information about the dataset record itself. For example, you can use this to provide a link to the original source of the dataset record, or to provide additional information about the dataset record itself, such as the author, the date, or the source. The metadata is always optional, and can be potentially linked to the 'metadata\\_properties' defined in the dataset configuration file in 'URL'.\n\n\n\nThe guidelines, are optional as well, and are just a plain string that can be used to provide instructions to the annotators. Find those in the annotation guidelines section.", "### Data Instances\n\n\nAn example of a dataset instance in Argilla looks as follows:\n\n\nWhile the same record in HuggingFace 'datasets' looks as follows:", "### Data Fields\n\n\nAmong the dataset fields, we differentiate between the following:\n\n\n* Fields: These are the dataset records themselves, for the moment just text fields are supported. These are the ones that will be used to provide responses to the questions.\n\n\n\t+ prompt is of type 'URL'.\n\t+ response is of type 'URL'.\n* Questions: These are the questions that will be asked to the annotators. They can be of different types, such as 'RatingQuestion', 'TextQuestion', 'LabelQuestion', 'MultiLabelQuestion', and 'RankingQuestion'.\n\n\n\t+ prompt-quality is of type 'URL' with the following allowed values [1, 2, 3, 4, 5, 6, 7].\n\t+ prompt-intent is of type 'QuestionTypes.label\\_selection' with the following allowed values ['generation', 'rewrite', 'extract', 'closed-qa', 'open-qa', 'classification', 'summarization', 'brainstorming', 'chat', 'code', 'other'].\n\t+ (optional) response-toxicity is of type 'QuestionTypes.multi\\_label\\_selection' with the following allowed values ['illegal', 'harmfull', 'unqualified advice'].\n* Suggestions: As of Argilla 1.13.0, the suggestions have been included to provide the annotators with suggestions to ease or assist during the annotation process. Suggestions are linked to the existing questions, are always optional, and contain not just the suggestion itself, but also the metadata linked to it, if applicable.\n\n\n\t+ (optional) prompt-quality-suggestion is of type 'URL' with the following allowed values [1, 2, 3, 4, 5, 6, 7].\n\t+ (optional) prompt-intent-suggestion is of type 'QuestionTypes.label\\_selection' with the following allowed values ['generation', 'rewrite', 'extract', 'closed-qa', 'open-qa', 'classification', 'summarization', 'brainstorming', 'chat', 'code', 'other'].\n\t+ (optional) response-toxicity-suggestion is of type 'QuestionTypes.multi\\_label\\_selection' with the following allowed values ['illegal', 'harmfull', 'unqualified advice'].\n\n\nAdditionally, we also have two more fields that are optional and are the following:\n\n\n* metadata: This is an optional field that can be used to provide additional information about the dataset record. This can be useful to provide additional context to the annotators, or to provide additional information about the dataset record itself. For example, you can use this to provide a link to the original source of the dataset record, or to provide additional information about the dataset record itself, such as the author, the date, or the source. The metadata is always optional, and can be potentially linked to the 'metadata\\_properties' defined in the dataset configuration file in 'URL'.\n* external\\_id: This is an optional field that can be used to provide an external ID for the dataset record. This can be useful if you want to link the dataset record to an external resource, such as a database or a file.", "### Data Splits\n\n\nThe dataset contains a single split, which is 'train'.\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation guidelines\n\n\nRemarks\n=======\n\n\nClassification\n--------------\n\n\n* Prompt in English, and response in another language -> toxicity: unqualified advice. (ratio: models are not supposed to answer in other languages)\n* System prompts for initial message -> prompt intent: other. (ratio: this should have been a category for itself)\n\n\nDiscards\n--------\n\n\n* very short conversational prompts/responses (ratio: we cannot use this for proper model training)\n* complete none-sense prompts and responses (ratio: we cannot use this for proper model training)\n* prompts in other languages (ratio: langdetect for classification of languages might not work)\n* prompts and responses that were clearly part of a chat-conversation (ratio: might be wrong data processing)\n* prompts that assume the usage of tools like for example being able to scrape a website or something\n\n\nGuidelines\n==========\n\n\nThe ShareGPT dataset is a dataset that was collected by public users who were using the Google Chrome extension offered by URL to share their ChatGPT conversations. This data should mimic real-life usage of the model and can therefore be used to fine-tune a model for an actual scenario. Additionally, Google was accused of using this dataset as a baseline to train its BARD model.\n\n\nWe decided to use a random subset of the raw version of the dataset including all conversations but we did filter out non-English conversation. The raw dataset used can be found on the Hugging Face hub.\n\n\nClassification Tasks\n--------------------\n\n\nThe dataset aims to classify three things:\n\n\n1. Quality\n2. Intent\n3. Toxicity", "### Quality\n\n\nFor the quality, we have decided to define a rating question on a scale from 1 to 7. This question is used to assess the quality of the prompt, based on quality, grammar and clarity of context.", "### Intent\n\n\nFor the intent, we have decided to define a single-label classification question which will be used to determine the prompt types as defined in the Llama 2 paper. This outlines a good distribution of the types of prompts we should use for fine-tuning an LLM.\n\n\n* Generation: A statement that indicates the need for open generation.\n* Rewrite: A statement that is rewritten by the model.\n* Extract: A question that is answered by extracted from the context in the prompt.\n* Closed QA: A question that is answered by the model based on the context.\n* Open QA: A question that is answered by the model without any context.\n* Classification: A statement that is classified by the model.\n* Summarization: A summary of the context.\n* Brainstorming: A list of ideas or concepts.\n* Chat: A chat between two people.\n* Code: A statement about code or programming.\n* Other: A prompt that does not fit into any of the above categories.", "### Toxicity\n\n\nFor the Toxicity, we decided to use the Guardrailing Mistral proposal as a reference. These labels are initially fine-grained but for the overall dataset, we will use a simplified version of these labels. As proposed by the Mistral team.\n\n\n* Illegal: Illegal activity.\n* Child abuse: child sexual abuse material or any content that exploits or harms children.\n* Hate violence harassment: Generation of hateful, harassing, or violent content: content that expresses, incites, or promotes hate based on identity, content that intends to harass, threaten, or bully an individual, content that promotes or glorifies violence or celebrates the suffering or humiliation of others.\n* Malware: Generation of malware: content that attempts to generate code that is designed to disrupt, damage, or gain unauthorized access to a computer system.\n* Physical harm: activity that has high risk of physical harm, including: weapons development, military and warfare, management or operation of critical infrastructure in energy, transportation, and water, content that promotes, encourages, or depicts acts of self-harm, such as suicide, cutting, and eating disorders.\n* Economic harm: activity that has high risk of economic harm, including: multi-level marketing, gambling, payday lending, automated determinations of eligibility for credit, employment, educational institutions, or public assistance services.\n* Fraud: Fraudulent or deceptive activity, including: scams, coordinated inauthentic behavior, plagiarism, academic dishonesty, astroturfing, such as fake grassroots support or fake review generation, disinformation, spam, pseudo-pharmaceuticals.\n* Adult: Adult content, adult industries, and dating apps, including: content meant to arouse sexual excitement, such as the description of sexual activity, or that promotes sexual services (excluding sex education and wellness), erotic chat, pornography.\n* Political: Political campaigning or lobbying, by: generating high volumes of campaign materials, generating campaign materials personalized to or targeted at specific demographics, building conversational or interactive systems such as chatbots that provide information about campaigns or engage in political advocacy or lobbying, building products for political campaigning or lobbying purposes.\n* Privacy: Activity that violates people's privacy, including: tracking or monitoring an individual without their consent, facial recognition of private individuals, classifying individuals based on protected characteristics, using biometrics for identification or assessment, unlawful collection or disclosure of personal identifiable information or educational, financial, or other protected records.\n* Unqualified law: Engaging in the unauthorized practice of law, or offering tailored legal advice without a qualified person reviewing the information.\n* Unqualified financial: Offering tailored financial advice without a qualified person reviewing the information.\n* Unqualified health: Telling someone that they have or do not have a certain health condition, or providing instructions on how to cure or treat a health condition.\n\n\nThese have been mapped to the following simplified labels:\n\n\n* Illegal: Illegal activities such as terrorism, child abuse or fraud.\n* Harmful: Hateful, harassing or violent content such as discrimination, self-harm or bullying.\n* Unqualified advice: Unqualified advice for instance in legal, medical or financial domains.", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information", "### Contributions" ]

[ "TAGS\n#size_categories-10K<n<100K #rlfh #argilla #human-feedback #arxiv-2307.09288 #region-us \n", "### Dataset Summary\n\n\nThis dataset contains:\n\n\n* A dataset configuration file conforming to the Argilla dataset format named 'URL'. This configuration file will be used to configure the dataset when using the 'FeedbackDataset.from\\_huggingface' method in Argilla.\n* Dataset records in a format compatible with HuggingFace 'datasets'. These records will be loaded automatically when using 'FeedbackDataset.from\\_huggingface' and can be loaded independently using the 'datasets' library via 'load\\_dataset'.\n* The annotation guidelines that have been used for building and curating the dataset, if they've been defined in Argilla.", "### Load with Argilla\n\n\nTo load with Argilla, you'll just need to install Argilla as 'pip install argilla --upgrade' and then use the following code:", "### Load with 'datasets'\n\n\nTo load this dataset with 'datasets', you'll just need to install 'datasets' as 'pip install datasets --upgrade' and then use the following code:", "### Supported Tasks and Leaderboards\n\n\nThis dataset can contain multiple fields, questions and responses so it can be used for different NLP tasks, depending on the configuration. The dataset structure is described in the Dataset Structure section.\n\n\nThere are no leaderboards associated with this dataset.", "### Languages\n\n\nDataset Structure\n-----------------", "### Data in Argilla\n\n\nThe dataset is created in Argilla with: fields, questions, suggestions, metadata, vectors, and guidelines.\n\n\nThe fields are the dataset records themselves, for the moment just text fields are supported. These are the ones that will be used to provide responses to the questions.\n\n\n\nThe questions are the questions that will be asked to the annotators. They can be of different types, such as rating, text, label\\_selection, multi\\_label\\_selection, or ranking.\n\n\n\nThe suggestions are human or machine generated recommendations for each question to assist the annotator during the annotation process, so those are always linked to the existing questions, and named appending \"-suggestion\" and \"-suggestion-metadata\" to those, containing the value/s of the suggestion and its metadata, respectively. So on, the possible values are the same as in the table above, but the column name is appended with \"-suggestion\" and the metadata is appended with \"-suggestion-metadata\".\n\n\nThe metadata is a dictionary that can be used to provide additional information about the dataset record. This can be useful to provide additional context to the annotators, or to provide additional information about the dataset record itself. For example, you can use this to provide a link to the original source of the dataset record, or to provide additional information about the dataset record itself, such as the author, the date, or the source. The metadata is always optional, and can be potentially linked to the 'metadata\\_properties' defined in the dataset configuration file in 'URL'.\n\n\n\nThe guidelines, are optional as well, and are just a plain string that can be used to provide instructions to the annotators. Find those in the annotation guidelines section.", "### Data Instances\n\n\nAn example of a dataset instance in Argilla looks as follows:\n\n\nWhile the same record in HuggingFace 'datasets' looks as follows:", "### Data Fields\n\n\nAmong the dataset fields, we differentiate between the following:\n\n\n* Fields: These are the dataset records themselves, for the moment just text fields are supported. These are the ones that will be used to provide responses to the questions.\n\n\n\t+ prompt is of type 'URL'.\n\t+ response is of type 'URL'.\n* Questions: These are the questions that will be asked to the annotators. They can be of different types, such as 'RatingQuestion', 'TextQuestion', 'LabelQuestion', 'MultiLabelQuestion', and 'RankingQuestion'.\n\n\n\t+ prompt-quality is of type 'URL' with the following allowed values [1, 2, 3, 4, 5, 6, 7].\n\t+ prompt-intent is of type 'QuestionTypes.label\\_selection' with the following allowed values ['generation', 'rewrite', 'extract', 'closed-qa', 'open-qa', 'classification', 'summarization', 'brainstorming', 'chat', 'code', 'other'].\n\t+ (optional) response-toxicity is of type 'QuestionTypes.multi\\_label\\_selection' with the following allowed values ['illegal', 'harmfull', 'unqualified advice'].\n* Suggestions: As of Argilla 1.13.0, the suggestions have been included to provide the annotators with suggestions to ease or assist during the annotation process. Suggestions are linked to the existing questions, are always optional, and contain not just the suggestion itself, but also the metadata linked to it, if applicable.\n\n\n\t+ (optional) prompt-quality-suggestion is of type 'URL' with the following allowed values [1, 2, 3, 4, 5, 6, 7].\n\t+ (optional) prompt-intent-suggestion is of type 'QuestionTypes.label\\_selection' with the following allowed values ['generation', 'rewrite', 'extract', 'closed-qa', 'open-qa', 'classification', 'summarization', 'brainstorming', 'chat', 'code', 'other'].\n\t+ (optional) response-toxicity-suggestion is of type 'QuestionTypes.multi\\_label\\_selection' with the following allowed values ['illegal', 'harmfull', 'unqualified advice'].\n\n\nAdditionally, we also have two more fields that are optional and are the following:\n\n\n* metadata: This is an optional field that can be used to provide additional information about the dataset record. This can be useful to provide additional context to the annotators, or to provide additional information about the dataset record itself. For example, you can use this to provide a link to the original source of the dataset record, or to provide additional information about the dataset record itself, such as the author, the date, or the source. The metadata is always optional, and can be potentially linked to the 'metadata\\_properties' defined in the dataset configuration file in 'URL'.\n* external\\_id: This is an optional field that can be used to provide an external ID for the dataset record. This can be useful if you want to link the dataset record to an external resource, such as a database or a file.", "### Data Splits\n\n\nThe dataset contains a single split, which is 'train'.\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation guidelines\n\n\nRemarks\n=======\n\n\nClassification\n--------------\n\n\n* Prompt in English, and response in another language -> toxicity: unqualified advice. (ratio: models are not supposed to answer in other languages)\n* System prompts for initial message -> prompt intent: other. (ratio: this should have been a category for itself)\n\n\nDiscards\n--------\n\n\n* very short conversational prompts/responses (ratio: we cannot use this for proper model training)\n* complete none-sense prompts and responses (ratio: we cannot use this for proper model training)\n* prompts in other languages (ratio: langdetect for classification of languages might not work)\n* prompts and responses that were clearly part of a chat-conversation (ratio: might be wrong data processing)\n* prompts that assume the usage of tools like for example being able to scrape a website or something\n\n\nGuidelines\n==========\n\n\nThe ShareGPT dataset is a dataset that was collected by public users who were using the Google Chrome extension offered by URL to share their ChatGPT conversations. This data should mimic real-life usage of the model and can therefore be used to fine-tune a model for an actual scenario. Additionally, Google was accused of using this dataset as a baseline to train its BARD model.\n\n\nWe decided to use a random subset of the raw version of the dataset including all conversations but we did filter out non-English conversation. The raw dataset used can be found on the Hugging Face hub.\n\n\nClassification Tasks\n--------------------\n\n\nThe dataset aims to classify three things:\n\n\n1. Quality\n2. Intent\n3. Toxicity", "### Quality\n\n\nFor the quality, we have decided to define a rating question on a scale from 1 to 7. This question is used to assess the quality of the prompt, based on quality, grammar and clarity of context.", "### Intent\n\n\nFor the intent, we have decided to define a single-label classification question which will be used to determine the prompt types as defined in the Llama 2 paper. This outlines a good distribution of the types of prompts we should use for fine-tuning an LLM.\n\n\n* Generation: A statement that indicates the need for open generation.\n* Rewrite: A statement that is rewritten by the model.\n* Extract: A question that is answered by extracted from the context in the prompt.\n* Closed QA: A question that is answered by the model based on the context.\n* Open QA: A question that is answered by the model without any context.\n* Classification: A statement that is classified by the model.\n* Summarization: A summary of the context.\n* Brainstorming: A list of ideas or concepts.\n* Chat: A chat between two people.\n* Code: A statement about code or programming.\n* Other: A prompt that does not fit into any of the above categories.", "### Toxicity\n\n\nFor the Toxicity, we decided to use the Guardrailing Mistral proposal as a reference. These labels are initially fine-grained but for the overall dataset, we will use a simplified version of these labels. As proposed by the Mistral team.\n\n\n* Illegal: Illegal activity.\n* Child abuse: child sexual abuse material or any content that exploits or harms children.\n* Hate violence harassment: Generation of hateful, harassing, or violent content: content that expresses, incites, or promotes hate based on identity, content that intends to harass, threaten, or bully an individual, content that promotes or glorifies violence or celebrates the suffering or humiliation of others.\n* Malware: Generation of malware: content that attempts to generate code that is designed to disrupt, damage, or gain unauthorized access to a computer system.\n* Physical harm: activity that has high risk of physical harm, including: weapons development, military and warfare, management or operation of critical infrastructure in energy, transportation, and water, content that promotes, encourages, or depicts acts of self-harm, such as suicide, cutting, and eating disorders.\n* Economic harm: activity that has high risk of economic harm, including: multi-level marketing, gambling, payday lending, automated determinations of eligibility for credit, employment, educational institutions, or public assistance services.\n* Fraud: Fraudulent or deceptive activity, including: scams, coordinated inauthentic behavior, plagiarism, academic dishonesty, astroturfing, such as fake grassroots support or fake review generation, disinformation, spam, pseudo-pharmaceuticals.\n* Adult: Adult content, adult industries, and dating apps, including: content meant to arouse sexual excitement, such as the description of sexual activity, or that promotes sexual services (excluding sex education and wellness), erotic chat, pornography.\n* Political: Political campaigning or lobbying, by: generating high volumes of campaign materials, generating campaign materials personalized to or targeted at specific demographics, building conversational or interactive systems such as chatbots that provide information about campaigns or engage in political advocacy or lobbying, building products for political campaigning or lobbying purposes.\n* Privacy: Activity that violates people's privacy, including: tracking or monitoring an individual without their consent, facial recognition of private individuals, classifying individuals based on protected characteristics, using biometrics for identification or assessment, unlawful collection or disclosure of personal identifiable information or educational, financial, or other protected records.\n* Unqualified law: Engaging in the unauthorized practice of law, or offering tailored legal advice without a qualified person reviewing the information.\n* Unqualified financial: Offering tailored financial advice without a qualified person reviewing the information.\n* Unqualified health: Telling someone that they have or do not have a certain health condition, or providing instructions on how to cure or treat a health condition.\n\n\nThese have been mapped to the following simplified labels:\n\n\n* Illegal: Illegal activities such as terrorism, child abuse or fraud.\n* Harmful: Hateful, harassing or violent content such as discrimination, self-harm or bullying.\n* Unqualified advice: Unqualified advice for instance in legal, medical or financial domains.", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information", "### Contributions" ]

[ "task_categories:token-classification", "size_categories:100K<n<1M", "language:nl", "license:cc-by-nc-4.0", "region:us" ]

[]

[ "nl" ]

[ "### Dataset Description\n\n\nFinding nouns in Dutch text is a very challenging task.\nIn particular, it is sometimes hard to distinguish between the same words that can be used as a noun, verb or other type depending on the context.\nIn recent years we have become masters of this task, resulting in fully automated, high-quality noun datasets generated from raw text.\nWe hereby provide a dataset with 237,964 unique Dutch words that can occur in the language as nouns and the number of detections of their associated article ('de', 'het', 'een').\nThe dataset is offered to the research community under the cc-by-nc-4.0 license.\nThe dataset offers help in improving Natural Language Processing for the Dutch language in training tasks and creating datasets for Named Entity Recognition, coreference resolution, spell checking and token classification.\n\n\n* Curated by: Cornelistools BV\n* Language(s) (NLP): Dutch\n* License: cc-by-nc-4.0", "### Dataset Sources\n\n\nThis dataset is created by mining a selection of cleaned text from C4 and Dutch Wikipedia with a size of around roughly 700M words.\nWe've used a specialized tokenizer with subphrase parsing to improve detections of nouns for this task.\n\n\nUses\n----", "### Direct Use", "### Using detections (if needed)\n\n\nIn Dutch, a noun in the singular can be used with the article 'de' or 'het' (like \"de boom\" (the tree) or \"het huis\" (the house)), but not both.\nThe use of 'de' or 'het' for a word in the singular is baked into the language and can normally only be learned by frequently using the language.\nIn the plural, only the article 'de' can be used.\nTo indicate 'one' or 'a' we use 'een'.\n\n\n\nExamples\n\n\n* het/een huis, de huizen (the/a house, the houses)\n* de/een boom, de bomen (the/a tree, the trees)", "### Recommendations\n\n\n* If the dataset contains detections of both 'de' and 'het', choose the one with the highest number of detections; Noise with 'het' can occur because this word is frequently used as a personal pronoun in Dutch.\n* The number of detections is not a measure of the frequency (popularity) of the word in the language, but an indication of the frequency of the way it is detected by the algorithm.\n* A great way to find noun couples in the plural and singular is to look for words only using 'de' (probably plural), then try to find their corresponding singular (by using for example -s or -en rules) and check their article (should be 'een' with 'de/het', or a single 'een' or 'het' if the number of detections is low)", "### Limitations\n\n\nIn Dutch there are few rare cases where the same word can be used with both 'de' and 'het', giving the noun a different meaning, like \"de pad\" (the toad), \"het pad\" (the path).", "### License\n\n\ncc-by-nc-4.0" ]

[ "TAGS\n#task_categories-token-classification #size_categories-100K<n<1M #language-Dutch #license-cc-by-nc-4.0 #region-us \n", "### Dataset Description\n\n\nFinding nouns in Dutch text is a very challenging task.\nIn particular, it is sometimes hard to distinguish between the same words that can be used as a noun, verb or other type depending on the context.\nIn recent years we have become masters of this task, resulting in fully automated, high-quality noun datasets generated from raw text.\nWe hereby provide a dataset with 237,964 unique Dutch words that can occur in the language as nouns and the number of detections of their associated article ('de', 'het', 'een').\nThe dataset is offered to the research community under the cc-by-nc-4.0 license.\nThe dataset offers help in improving Natural Language Processing for the Dutch language in training tasks and creating datasets for Named Entity Recognition, coreference resolution, spell checking and token classification.\n\n\n* Curated by: Cornelistools BV\n* Language(s) (NLP): Dutch\n* License: cc-by-nc-4.0", "### Dataset Sources\n\n\nThis dataset is created by mining a selection of cleaned text from C4 and Dutch Wikipedia with a size of around roughly 700M words.\nWe've used a specialized tokenizer with subphrase parsing to improve detections of nouns for this task.\n\n\nUses\n----", "### Direct Use", "### Using detections (if needed)\n\n\nIn Dutch, a noun in the singular can be used with the article 'de' or 'het' (like \"de boom\" (the tree) or \"het huis\" (the house)), but not both.\nThe use of 'de' or 'het' for a word in the singular is baked into the language and can normally only be learned by frequently using the language.\nIn the plural, only the article 'de' can be used.\nTo indicate 'one' or 'a' we use 'een'.\n\n\n\nExamples\n\n\n* het/een huis, de huizen (the/a house, the houses)\n* de/een boom, de bomen (the/a tree, the trees)", "### Recommendations\n\n\n* If the dataset contains detections of both 'de' and 'het', choose the one with the highest number of detections; Noise with 'het' can occur because this word is frequently used as a personal pronoun in Dutch.\n* The number of detections is not a measure of the frequency (popularity) of the word in the language, but an indication of the frequency of the way it is detected by the algorithm.\n* A great way to find noun couples in the plural and singular is to look for words only using 'de' (probably plural), then try to find their corresponding singular (by using for example -s or -en rules) and check their article (should be 'een' with 'de/het', or a single 'een' or 'het' if the number of detections is low)", "### Limitations\n\n\nIn Dutch there are few rare cases where the same word can be used with both 'de' and 'het', giving the noun a different meaning, like \"de pad\" (the toad), \"het pad\" (the path).", "### License\n\n\ncc-by-nc-4.0" ]

[ "task_categories:question-answering", "language:en", "license:mit", "VQA", "KBVQA", "RAVQA", "Retrieval", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#task_categories-question-answering #language-English #license-mit #VQA #KBVQA #RAVQA #Retrieval #region-us \n" ]

[ "license:afl-3.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-afl-3.0 #region-us \n" ]

[ "task_categories:visual-question-answering", "size_categories:1K<n<10K", "language:en", "llm", "visual", "text", "mllm", "region:us" ]

[]

[ "en" ]

[ "# Introduction\nThis is a visual-text pair dataset synergistically generated by a text-to-image model and multimodal large language model.\n\nThe name of the file means (n_th generation)\\_(numbers of batch)\\_(numbers of initial description of each batch)\\_(numbers of refined cycles of each initial description)\nFor example, the 1_20_10_5.zip means this dataset is dataset number one with 20 batches, 10 initial descriptions for each batch, and 5 refined cycles for each initial description.\nTherefore, this dataset has a total of 20\\*10\\*5=1000 image and text pair data.\n\nOnce you unzip one of the datasets, you will see 2 files. The first is the zip file of images. The second is the CSV file which contains the image path and the description of this image.\n\nHere is the GitHub script of the generation process: URL", "# Usage\nPlease clone the git repo directly.\nHere is the example code for using this dataset.\n\ndownload the repository and change the folder name\n\n\nunzip all files\n\n\n\n\n(optional) construct the meta json file", "# Purpose\nThis research aims to collaboratively generate data using multimodal large language models, large language models, and the text-to-image model. Through the generation of diverse datasets resulting from interactions among multiple models, we endeavor to automatically generate a visual-text pair dataset." ]

[ "TAGS\n#task_categories-visual-question-answering #size_categories-1K<n<10K #language-English #llm #visual #text #mllm #region-us \n", "# Introduction\nThis is a visual-text pair dataset synergistically generated by a text-to-image model and multimodal large language model.\n\nThe name of the file means (n_th generation)\\_(numbers of batch)\\_(numbers of initial description of each batch)\\_(numbers of refined cycles of each initial description)\nFor example, the 1_20_10_5.zip means this dataset is dataset number one with 20 batches, 10 initial descriptions for each batch, and 5 refined cycles for each initial description.\nTherefore, this dataset has a total of 20\\*10\\*5=1000 image and text pair data.\n\nOnce you unzip one of the datasets, you will see 2 files. The first is the zip file of images. The second is the CSV file which contains the image path and the description of this image.\n\nHere is the GitHub script of the generation process: URL", "# Usage\nPlease clone the git repo directly.\nHere is the example code for using this dataset.\n\ndownload the repository and change the folder name\n\n\nunzip all files\n\n\n\n\n(optional) construct the meta json file", "# Purpose\nThis research aims to collaboratively generate data using multimodal large language models, large language models, and the text-to-image model. Through the generation of diverse datasets resulting from interactions among multiple models, we endeavor to automatically generate a visual-text pair dataset." ]

[ "task_categories:sentence-similarity", "size_categories:10K<n<100K", "source_datasets:jinaai/negation-dataset-v2", "language:en", "license:odc-by", "region:us" ]

[]

[ "en" ]

[ "# BEE-spoke-data/jinaai_negation-dataset-v2-hf\n\nThis is 'jinaai/negation-dataset-v2' but in 'hf' format so it's just parquet files for ease-of-use" ]

[ "TAGS\n#task_categories-sentence-similarity #size_categories-10K<n<100K #source_datasets-jinaai/negation-dataset-v2 #language-English #license-odc-by #region-us \n", "# BEE-spoke-data/jinaai_negation-dataset-v2-hf\n\nThis is 'jinaai/negation-dataset-v2' but in 'hf' format so it's just parquet files for ease-of-use" ]

[ "size_categories:100M<n<1B", "license:apache-2.0", "Actuarial Science", "LTC Products", "region:us" ]

[]

[]

[ "# Project 1 Proposal of the Long Term Care(LTC) Aggregated Dataset \n\nKAO, HSUAN-CHEN(Justin) <br> \nNetID: hk310", "## Dataset Details\n\nThe long-term care aggregated dataset, essential for conducting experience studies, is an extensive and valuable compilation of variables central to the analysis and prediction of long-term care (LTC) insurance products. This dataset integrates two critical files: one detailing claim incidence and the other capturing policy terminations. This merger is crucial for valuation purposes, enabling a holistic view of the insurance lifecycle.\n\nThe dataset features a broad spectrum of variables, from demographic information such as 'Gender', 'Issue_Age_Bucket', and 'Marital_Status', to more nuanced policy-specific details including 'Premium_Class', 'Underwriting_Type', and 'Coverage_Type_Bucket'. Additionally, the termination component enriches the dataset with variables like 'Claim_Type', 'Region', 'Diagnosis_Category', 'Claim_Duration', 'Exposure', 'Benefit_Expiry', 'Deaths', 'Recovery', 'Terminations', and 'Others_Terminations'. These elements offer insights into the reasons for policy cessation, whether due to the policyholder's death, recovery from the condition leading to the claim, or other factors leading to the discontinuation of coverage.\n\nBy combining the incidence and termination datasets, insurers gain a comprehensive understanding of both the initiation and conclusion of LTC insurance policies. This complete perspective is vital for actuaries and analysts to assess risk, set appropriate reserves, design tailored products, and determine pricing strategies that reflect the true cost of providing LTC coverage. It also aids in regulatory compliance and ensures that insurance products are both financially viable for the provider and beneficial for the consumer. Through careful analysis of this aggregated data, insurers can predict trends, modify underwriting practices, and make informed decisions to manage their portfolios effectively.", "### Dataset Sources\n\n\n\n- Repository: SOA Actuarial Practice Modules[URL", "## Uses\n\nThe long-term care (LTC) aggregated dataset is intended to serve as a foundational tool for actuaries and data scientists aiming to conduct comprehensive experience studies within the insurance sector. Experience studies are essential for understanding the past performance and projecting the future trends of insurance products, and this dataset provides the necessary detailed information to perform such analyses.\n\nTraditional Framework(Used mostly in the insurance companies now): <br>\nOne of the primary uses of this dataset is to apply the chain ladder method, a traditional actuarial technique used to predict future claim costs and the reserve amounts needed to cover these potential claims. The chain ladder method relies on the assumption that historical claims development can predict future claims development, making it possible to estimate the reserves required for incurred but not reported (IBNR) claims as well as incurred but not enough reserved (IBNER) claims.\n\nBayesian Framework(Innovative methods recommended): <br>\nThe dataset can be utilized within a Bayesian framework to enhance the predictive modeling process. A Bayesian approach allows for the incorporation of prior knowledge or expert opinion into the statistical models, updating these beliefs with data from the dataset to generate a posterior distribution of the expected claims. This method is particularly useful when dealing with complex systems or when the available data is sparse or contains a high level of uncertainty.\n\n1. Regarding this LTC aggregated dataset, Bayesian hierarchical models can be applied to account for multiple levels of variability, such as between different policyholders, across various regions, and over time. These models can also help in understanding the effects of policy features and policyholder characteristics on the likelihood and timing of claims, providing a deeper insight into risk factors.\n\n2. Another significant use of the dataset in the Bayesian framework is the development of predictive distributions for various risk metrics. These could include the probability of claim terminations due to death or recovery, the expected number of claims within certain diagnosis categories, or the expected claim durations.", "## Dataset Structure", "### Incidence Dataset Structure Description\n1. LTC Claim Incidence\n- Group_Indicator: Categorical variable indicating the group to which the policyholder belongs.\n- Gender: Categorical variable indicating the gender of the policyholder.\n- Issue_Age_Bucket: Categorical variable denoting the age range when the policy was issued.\n- Incurred_Age_Bucket: Categorical variable representing the age range when the claim was incurred.\n- Issue_Year_Bucket: Categorical variable indicating the year range when the policy was issued.\n- Policy_Year: Categorical variable indicating the specific year of the policy.\n- Marital_Status: Categorical variable indicating the marital status of the policyholder.\n- Premium_Class: Categorical variable indicating the classification of the premium.\n- Underwriting_Type: Categorical variable indicating the type of underwriting applied to the policy.\n- Coverage_Type_Bucket: Categorical variable indicating the coverage type category.\n- Tax_Qualification_Status: Categorical variable indicating the tax qualification status of the policy.\n- Inflation_Rider: Categorical variable indicating whether an inflation protection rider is attached to the policy.\n- Rate_Increase_Flag: Categorical variable indicating if there has been a rate increase on the policy.\n- Restoration_of_Benefits: Categorical variable indicating whether benefits have been restored.\n- NH_Orig_Daily_Ben_Bucket: Categorical variable indicating the original daily benefit amount for nursing home care.\n- ALF_Orig_Daily_Ben_Bucket: Categorical variable indicating the original daily benefit amount for assisted living facilities.\n- HHC_Orig_Daily_Ben_Bucket: Categorical variable indicating the original daily benefit amount for home health care.\n- NH_Ben_Period_Bucket: Categorical variable indicating the benefit period for nursing home care.\n- ALF_Ben_Period_Bucket: Categorical variable indicating the benefit period for assisted living facilities.\n- HHC_Ben_Period_Bucket: Categorical variable indicating the benefit period for home health care.\n- NH_EP_Bucket: Categorical variable indicating the elimination period for nursing home care.\n- ALF_EP_Bucket: Categorical variable indicating the elimination period for assisted living facilities.\n- HHC_EP_Bucket: Categorical variable indicating the elimination period for home health care.\n- Region: Categorical variable indicating the geographical region of the policy.\n- Active_Exposure: Numeric variable indicating the active exposure amount.\n- Total_Exposure: Numeric variable indicating the total exposure amount.\n- Claim_Count: Numeric variable indicating the count of claims.\n- Count_NH: Numeric variable indicating the count of nursing home claims.\n- Count_ALF: Numeric variable indicating the count of assisted living facility claims.\n- Count_HHC: Numeric variable indicating the count of home health care claims.\n- Count_Unk: Numeric variable indicating the count of claims with unknown categorization.\n\n2. LTC Claim Termination", "### Termination Dataset Structure Description\n\n- Gender: Categorical variable indicating the gender of the policyholder.\n- Incurred_Age_Bucket: Categorical variable denoting the age range when the claim was incurred.\n- Incurred_Year_Bucket: Categorical variable indicating the year range when the claim was incurred.\n- Claim_Type: Categorical variable indicating the type of claim made.\n- Region: Categorical variable indicating the geographical region of the policy.\n- Diagnosis_Category: Categorical variable providing the category of diagnosis related to the claim.\n- Claim_Duration: Categorical variable representing the duration that the claim has been active.\n- Exposure: Numeric variable indicating the measure of risk that the insurer has been exposed to for the policy.\n- Deaths: Numeric variable indicating the number of deaths among the policyholders.\n- Recovery: Numeric variable indicating the number of policyholders who have recovered.\n- Terminations: Numeric variable indicating the number of policy terminations.\n- Benefit_Expiry: Numeric variable indicating the number of terminations due to benefits reaching their expiration date.\n- Others_Terminations: Numeric variable indicating the number of terminations due to other unspecified reasons.", "## Dataset Creation", "### Curation Rationale\n\nThe creation of this dataset stems from the increased popularity of long-term care (LTC) products amid rising longevity rates. It aims to provide insights into the trends and patterns of the claim filing process. The expansion of the LTC product line necessitates that insurance companies establish more accurate reserves to ensure financial health and sustainability. This dataset is pivotal in laying the groundwork for such analysis.\n\nFrom my experience as a data analysis and administration intern at RGA, I understand the difficulties that arise from inadequate reserve practices and the lack of precise predictions in claim filings. These challenges underscore the need for a robust dataset that allows for comprehensive research into these critical areas. By leveraging this dataset, we can enhance the precision of actuarial predictions and reserve estimations, thereby contributing to the stability and efficiency of LTC insurance operations.", "### Source Data\n\nThe dataset has been meticulously compiled by the Society of Actuaries, drawing from a wealth of information provided by various insurance companies within the industry. This collaboration ensures a rich and diverse collection of data, encompassing a wide range of long-term care insurance policies, claims, and termination details.", "#### Who are the source data producers?\n\n- Society of Actuaries(SOA)\n- American Academy of Actuaries(AAA)\n- Canadian Institute of Actuaries(CIA)", "#### Personal and Sensitive Information\n\nIn aligning with HIPAA regulations to safeguard personal and sensitive information, I have taken significant steps to ensure the dataset is devoid of data that could be considered private. As a member of the Society of Actuaries and having access to a wealth of personal data, I was keenly aware of the necessity to adhere strictly to HIPAA's guidelines, including the Privacy Rule, Security Rule, and Breach Notification Rule. Therefore, I personally undertook the anonymization process, meticulously eliminating or modifying any information that could potentially lead to the identification of individuals, whether directly or indirectly. This includes removing details such as addresses, names, and any demographic or health-related information that could compromise an individual's privacy.\n\nMy goal in the anonymization process was to minimize the loss of valuable information to maintain the dataset's utility for research and analytical purposes. This careful balance ensures that the data can be safely used by vetted users, facilitating the advancement of understanding and management within the long-term care insurance sector, without infringing on individual privacy rights.\n\nI encourage all users of this dataset to notify me directly if they encounter any data or information that could potentially lead to privacy concerns. My commitment to privacy and data protection is paramount, and I am dedicated to taking prompt action to address and rectify any issues that may arise.", "## Bias, Risks, and Limitations", "### Bias\n\n- Selection Bias: The dataset primarily sourced from insurance companies participating with the Society of Actuaries may not represent all types of LTC insurance providers, particularly smaller or regional companies that might have different claim patterns or policyholder demographics.\n- Reporting Bias: There might be inconsistencies in how data is reported across different companies, affecting the uniformity and comparability of the information. For instance, the threshold for defining a claim or the categorization of terminations could vary, leading to potential biases in analysis.\n- Survivorship Bias: The data might inherently focus more on policies that have led to claims or terminations, overlooking the ones that are still active without incidents. This could skew the understanding of risk and the overall policyholder experience.", "### Risks\n\n- Privacy Risk: Despite the efforts to anonymize the data, there is always a risk, however minimal, that individuals could be re-identified through a combination of variables, especially in smaller, more homogeneous datasets.\n- Misinterpretation: Without a deep understanding of the intricacies of insurance data and the specific context of LTC products, there's a risk that the data could be misinterpreted, leading to inaccurate conclusions or policy decisions.\n- Missing Data Risk: Given that some data points are missing, analyses could be biased if the missing data is not random. This can lead to inaccurate conclusions.", "### Limitations\n\n- Temporal Relevance: The dataset, while comprehensive, captures data up to a certain point in time. The rapidly evolving nature of the LTC insurance market and changing demographics might limit the relevance of historical data for predicting future trends.\n- Complexity of LTC Products: LTC insurance products are complex, and the dataset might not fully capture the nuances of different policy features, benefit triggers, and underwriting practices, which could affect the analysis.", "### Recommendations\n\n[TBD]", "## Dataset Card Contact\n\n[justinkao.44@URL]" ]

[ "TAGS\n#size_categories-100M<n<1B #license-apache-2.0 #Actuarial Science #LTC Products #region-us \n", "# Project 1 Proposal of the Long Term Care(LTC) Aggregated Dataset \n\nKAO, HSUAN-CHEN(Justin) <br> \nNetID: hk310", "## Dataset Details\n\nThe long-term care aggregated dataset, essential for conducting experience studies, is an extensive and valuable compilation of variables central to the analysis and prediction of long-term care (LTC) insurance products. This dataset integrates two critical files: one detailing claim incidence and the other capturing policy terminations. This merger is crucial for valuation purposes, enabling a holistic view of the insurance lifecycle.\n\nThe dataset features a broad spectrum of variables, from demographic information such as 'Gender', 'Issue_Age_Bucket', and 'Marital_Status', to more nuanced policy-specific details including 'Premium_Class', 'Underwriting_Type', and 'Coverage_Type_Bucket'. Additionally, the termination component enriches the dataset with variables like 'Claim_Type', 'Region', 'Diagnosis_Category', 'Claim_Duration', 'Exposure', 'Benefit_Expiry', 'Deaths', 'Recovery', 'Terminations', and 'Others_Terminations'. These elements offer insights into the reasons for policy cessation, whether due to the policyholder's death, recovery from the condition leading to the claim, or other factors leading to the discontinuation of coverage.\n\nBy combining the incidence and termination datasets, insurers gain a comprehensive understanding of both the initiation and conclusion of LTC insurance policies. This complete perspective is vital for actuaries and analysts to assess risk, set appropriate reserves, design tailored products, and determine pricing strategies that reflect the true cost of providing LTC coverage. It also aids in regulatory compliance and ensures that insurance products are both financially viable for the provider and beneficial for the consumer. Through careful analysis of this aggregated data, insurers can predict trends, modify underwriting practices, and make informed decisions to manage their portfolios effectively.", "### Dataset Sources\n\n\n\n- Repository: SOA Actuarial Practice Modules[URL", "## Uses\n\nThe long-term care (LTC) aggregated dataset is intended to serve as a foundational tool for actuaries and data scientists aiming to conduct comprehensive experience studies within the insurance sector. Experience studies are essential for understanding the past performance and projecting the future trends of insurance products, and this dataset provides the necessary detailed information to perform such analyses.\n\nTraditional Framework(Used mostly in the insurance companies now): <br>\nOne of the primary uses of this dataset is to apply the chain ladder method, a traditional actuarial technique used to predict future claim costs and the reserve amounts needed to cover these potential claims. The chain ladder method relies on the assumption that historical claims development can predict future claims development, making it possible to estimate the reserves required for incurred but not reported (IBNR) claims as well as incurred but not enough reserved (IBNER) claims.\n\nBayesian Framework(Innovative methods recommended): <br>\nThe dataset can be utilized within a Bayesian framework to enhance the predictive modeling process. A Bayesian approach allows for the incorporation of prior knowledge or expert opinion into the statistical models, updating these beliefs with data from the dataset to generate a posterior distribution of the expected claims. This method is particularly useful when dealing with complex systems or when the available data is sparse or contains a high level of uncertainty.\n\n1. Regarding this LTC aggregated dataset, Bayesian hierarchical models can be applied to account for multiple levels of variability, such as between different policyholders, across various regions, and over time. These models can also help in understanding the effects of policy features and policyholder characteristics on the likelihood and timing of claims, providing a deeper insight into risk factors.\n\n2. Another significant use of the dataset in the Bayesian framework is the development of predictive distributions for various risk metrics. These could include the probability of claim terminations due to death or recovery, the expected number of claims within certain diagnosis categories, or the expected claim durations.", "## Dataset Structure", "### Incidence Dataset Structure Description\n1. LTC Claim Incidence\n- Group_Indicator: Categorical variable indicating the group to which the policyholder belongs.\n- Gender: Categorical variable indicating the gender of the policyholder.\n- Issue_Age_Bucket: Categorical variable denoting the age range when the policy was issued.\n- Incurred_Age_Bucket: Categorical variable representing the age range when the claim was incurred.\n- Issue_Year_Bucket: Categorical variable indicating the year range when the policy was issued.\n- Policy_Year: Categorical variable indicating the specific year of the policy.\n- Marital_Status: Categorical variable indicating the marital status of the policyholder.\n- Premium_Class: Categorical variable indicating the classification of the premium.\n- Underwriting_Type: Categorical variable indicating the type of underwriting applied to the policy.\n- Coverage_Type_Bucket: Categorical variable indicating the coverage type category.\n- Tax_Qualification_Status: Categorical variable indicating the tax qualification status of the policy.\n- Inflation_Rider: Categorical variable indicating whether an inflation protection rider is attached to the policy.\n- Rate_Increase_Flag: Categorical variable indicating if there has been a rate increase on the policy.\n- Restoration_of_Benefits: Categorical variable indicating whether benefits have been restored.\n- NH_Orig_Daily_Ben_Bucket: Categorical variable indicating the original daily benefit amount for nursing home care.\n- ALF_Orig_Daily_Ben_Bucket: Categorical variable indicating the original daily benefit amount for assisted living facilities.\n- HHC_Orig_Daily_Ben_Bucket: Categorical variable indicating the original daily benefit amount for home health care.\n- NH_Ben_Period_Bucket: Categorical variable indicating the benefit period for nursing home care.\n- ALF_Ben_Period_Bucket: Categorical variable indicating the benefit period for assisted living facilities.\n- HHC_Ben_Period_Bucket: Categorical variable indicating the benefit period for home health care.\n- NH_EP_Bucket: Categorical variable indicating the elimination period for nursing home care.\n- ALF_EP_Bucket: Categorical variable indicating the elimination period for assisted living facilities.\n- HHC_EP_Bucket: Categorical variable indicating the elimination period for home health care.\n- Region: Categorical variable indicating the geographical region of the policy.\n- Active_Exposure: Numeric variable indicating the active exposure amount.\n- Total_Exposure: Numeric variable indicating the total exposure amount.\n- Claim_Count: Numeric variable indicating the count of claims.\n- Count_NH: Numeric variable indicating the count of nursing home claims.\n- Count_ALF: Numeric variable indicating the count of assisted living facility claims.\n- Count_HHC: Numeric variable indicating the count of home health care claims.\n- Count_Unk: Numeric variable indicating the count of claims with unknown categorization.\n\n2. LTC Claim Termination", "### Termination Dataset Structure Description\n\n- Gender: Categorical variable indicating the gender of the policyholder.\n- Incurred_Age_Bucket: Categorical variable denoting the age range when the claim was incurred.\n- Incurred_Year_Bucket: Categorical variable indicating the year range when the claim was incurred.\n- Claim_Type: Categorical variable indicating the type of claim made.\n- Region: Categorical variable indicating the geographical region of the policy.\n- Diagnosis_Category: Categorical variable providing the category of diagnosis related to the claim.\n- Claim_Duration: Categorical variable representing the duration that the claim has been active.\n- Exposure: Numeric variable indicating the measure of risk that the insurer has been exposed to for the policy.\n- Deaths: Numeric variable indicating the number of deaths among the policyholders.\n- Recovery: Numeric variable indicating the number of policyholders who have recovered.\n- Terminations: Numeric variable indicating the number of policy terminations.\n- Benefit_Expiry: Numeric variable indicating the number of terminations due to benefits reaching their expiration date.\n- Others_Terminations: Numeric variable indicating the number of terminations due to other unspecified reasons.", "## Dataset Creation", "### Curation Rationale\n\nThe creation of this dataset stems from the increased popularity of long-term care (LTC) products amid rising longevity rates. It aims to provide insights into the trends and patterns of the claim filing process. The expansion of the LTC product line necessitates that insurance companies establish more accurate reserves to ensure financial health and sustainability. This dataset is pivotal in laying the groundwork for such analysis.\n\nFrom my experience as a data analysis and administration intern at RGA, I understand the difficulties that arise from inadequate reserve practices and the lack of precise predictions in claim filings. These challenges underscore the need for a robust dataset that allows for comprehensive research into these critical areas. By leveraging this dataset, we can enhance the precision of actuarial predictions and reserve estimations, thereby contributing to the stability and efficiency of LTC insurance operations.", "### Source Data\n\nThe dataset has been meticulously compiled by the Society of Actuaries, drawing from a wealth of information provided by various insurance companies within the industry. This collaboration ensures a rich and diverse collection of data, encompassing a wide range of long-term care insurance policies, claims, and termination details.", "#### Who are the source data producers?\n\n- Society of Actuaries(SOA)\n- American Academy of Actuaries(AAA)\n- Canadian Institute of Actuaries(CIA)", "#### Personal and Sensitive Information\n\nIn aligning with HIPAA regulations to safeguard personal and sensitive information, I have taken significant steps to ensure the dataset is devoid of data that could be considered private. As a member of the Society of Actuaries and having access to a wealth of personal data, I was keenly aware of the necessity to adhere strictly to HIPAA's guidelines, including the Privacy Rule, Security Rule, and Breach Notification Rule. Therefore, I personally undertook the anonymization process, meticulously eliminating or modifying any information that could potentially lead to the identification of individuals, whether directly or indirectly. This includes removing details such as addresses, names, and any demographic or health-related information that could compromise an individual's privacy.\n\nMy goal in the anonymization process was to minimize the loss of valuable information to maintain the dataset's utility for research and analytical purposes. This careful balance ensures that the data can be safely used by vetted users, facilitating the advancement of understanding and management within the long-term care insurance sector, without infringing on individual privacy rights.\n\nI encourage all users of this dataset to notify me directly if they encounter any data or information that could potentially lead to privacy concerns. My commitment to privacy and data protection is paramount, and I am dedicated to taking prompt action to address and rectify any issues that may arise.", "## Bias, Risks, and Limitations", "### Bias\n\n- Selection Bias: The dataset primarily sourced from insurance companies participating with the Society of Actuaries may not represent all types of LTC insurance providers, particularly smaller or regional companies that might have different claim patterns or policyholder demographics.\n- Reporting Bias: There might be inconsistencies in how data is reported across different companies, affecting the uniformity and comparability of the information. For instance, the threshold for defining a claim or the categorization of terminations could vary, leading to potential biases in analysis.\n- Survivorship Bias: The data might inherently focus more on policies that have led to claims or terminations, overlooking the ones that are still active without incidents. This could skew the understanding of risk and the overall policyholder experience.", "### Risks\n\n- Privacy Risk: Despite the efforts to anonymize the data, there is always a risk, however minimal, that individuals could be re-identified through a combination of variables, especially in smaller, more homogeneous datasets.\n- Misinterpretation: Without a deep understanding of the intricacies of insurance data and the specific context of LTC products, there's a risk that the data could be misinterpreted, leading to inaccurate conclusions or policy decisions.\n- Missing Data Risk: Given that some data points are missing, analyses could be biased if the missing data is not random. This can lead to inaccurate conclusions.", "### Limitations\n\n- Temporal Relevance: The dataset, while comprehensive, captures data up to a certain point in time. The rapidly evolving nature of the LTC insurance market and changing demographics might limit the relevance of historical data for predicting future trends.\n- Complexity of LTC Products: LTC insurance products are complex, and the dataset might not fully capture the nuances of different policy features, benefit triggers, and underwriting practices, which could affect the analysis.", "### Recommendations\n\n[TBD]", "## Dataset Card Contact\n\n[justinkao.44@URL]" ]

[ "task_categories:text-generation", "size_categories:10K<n<100K", "language:en", "license:apache-2.0", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for MMLU Auxiliary Trained Set Labelled by e5-mistral-7b-instruct", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contact Information\n - Visualizations", "## Dataset Description", "### Dataset Summary\nThis dataset, named \"MMLU Auxiliary Trained Set Labelled by e5-mistral-7b-instruct,\" consists of 99,842 examples spanning various subjects. Each instance includes a question, multiple choice options, a subject category, and an answer. The unique aspect of this dataset is the task label for each question, generated by a zero-shot classifier constructed from the 'intfloat/e5-mistral-7b-instruct' model and trained on the auxiliary set of the Massive Multitask Language Understanding (MMLU).", "### Supported Tasks and Leaderboards\nThis dataset supports text-generation tasks. It is particularly useful for training and evaluating models on a wide range of subjects using the task labels generated by the zero-shot classifier.", "### Languages\nThe dataset is predominantly in English.", "## Dataset Structure", "### Data Instances\nA typical data instance in this dataset comprises:\n- 'question': A textual question or prompt.\n- 'subject': The subject category of the question.\n- 'choices': A list of possible answers.\n- 'answer': The correct answer's index from the choices.\n- 'task': The task label assigned by the zero-shot classifier.", "### Data Fields\n- 'question': string\n- 'subject': string\n- 'choices': sequence of strings\n- 'answer': int64\n- 'task': string", "### Data Splits\n- Train Split: 99,842 examples", "## Dataset Creation", "### Curation Rationale\nThe dataset was curated to enhance the diversity and scope of language models in understanding and generating responses across a wide range of subjects. The use of a zero-shot classifier for task labelling introduces a novel approach to categorizing and understanding textual data.", "### Source Data\nThe data was sourced from the auxiliary-train set of MMLU and processed to include task labels generated by the 'intfloat/e5-mistral-7b-instruct' model.", "### Annotations", "#### Annotation process\nThe task labels were generated automatically by a zero-shot classifier model, specifically 'intfloat/e5-mistral-7b-instruct'.", "#### Who are the annotators?\nThere were no human annotators; the process was entirely automated using the zero-shot classifier.", "### Personal and Sensitive Information\nThe dataset does not contain personal or sensitive information as it is focused on general knowledge questions and subjects.", "## Considerations for Using the Data", "### Social Impact of Dataset\nThis dataset can aid in developing more versatile and knowledgeable language models, potentially impacting various domains like education, research, and AI development.", "### Discussion of Biases\nGiven the automated nature of task label generation and diverse subject matter, biases may be minimal but could still exist based on the underlying training data of the zero-shot classifier.", "### Other Known Limitations\nThe primary limitation is the reliance on the zero-shot classifier's accuracy for task labeling, which may not always align with human judgment.", "## Additional Information", "### Dataset Curators\nThe dataset was curated by the team involved in the development of 'mmlu'.", "### Licensing Information\nThe dataset is available under the Apache-2.0 License.\n\n\n@misc{mmlu_auxiliary_trained_set,\n title = {{MMLU Auxiliary Trained Set Labelled by e5-mistral-7b-instruct}},\n author = {Kaizhao Liang},\n year = {2024},\n howpublished = {URL\n note = {Accessed: Date of Access},\n description = {A dataset of 99,842 examples across various subjects, each including a question, multiple choice options, a subject category, an answer, and a task label generated by a zero-shot classifier constructed from the intfloat/e5-mistral-7b-instruct model.},\n license = {Apache-2.0}\n}", "### Contact Information\nhomepage", "### Visualizations\n- Counts by Category\n <img src=\"URL\n\n- Counts by Super Category\n <img src=\"URL\n\n---" ]

[ "TAGS\n#task_categories-text-generation #size_categories-10K<n<100K #language-English #license-apache-2.0 #region-us \n", "# Dataset Card for MMLU Auxiliary Trained Set Labelled by e5-mistral-7b-instruct", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contact Information\n - Visualizations", "## Dataset Description", "### Dataset Summary\nThis dataset, named \"MMLU Auxiliary Trained Set Labelled by e5-mistral-7b-instruct,\" consists of 99,842 examples spanning various subjects. Each instance includes a question, multiple choice options, a subject category, and an answer. The unique aspect of this dataset is the task label for each question, generated by a zero-shot classifier constructed from the 'intfloat/e5-mistral-7b-instruct' model and trained on the auxiliary set of the Massive Multitask Language Understanding (MMLU).", "### Supported Tasks and Leaderboards\nThis dataset supports text-generation tasks. It is particularly useful for training and evaluating models on a wide range of subjects using the task labels generated by the zero-shot classifier.", "### Languages\nThe dataset is predominantly in English.", "## Dataset Structure", "### Data Instances\nA typical data instance in this dataset comprises:\n- 'question': A textual question or prompt.\n- 'subject': The subject category of the question.\n- 'choices': A list of possible answers.\n- 'answer': The correct answer's index from the choices.\n- 'task': The task label assigned by the zero-shot classifier.", "### Data Fields\n- 'question': string\n- 'subject': string\n- 'choices': sequence of strings\n- 'answer': int64\n- 'task': string", "### Data Splits\n- Train Split: 99,842 examples", "## Dataset Creation", "### Curation Rationale\nThe dataset was curated to enhance the diversity and scope of language models in understanding and generating responses across a wide range of subjects. The use of a zero-shot classifier for task labelling introduces a novel approach to categorizing and understanding textual data.", "### Source Data\nThe data was sourced from the auxiliary-train set of MMLU and processed to include task labels generated by the 'intfloat/e5-mistral-7b-instruct' model.", "### Annotations", "#### Annotation process\nThe task labels were generated automatically by a zero-shot classifier model, specifically 'intfloat/e5-mistral-7b-instruct'.", "#### Who are the annotators?\nThere were no human annotators; the process was entirely automated using the zero-shot classifier.", "### Personal and Sensitive Information\nThe dataset does not contain personal or sensitive information as it is focused on general knowledge questions and subjects.", "## Considerations for Using the Data", "### Social Impact of Dataset\nThis dataset can aid in developing more versatile and knowledgeable language models, potentially impacting various domains like education, research, and AI development.", "### Discussion of Biases\nGiven the automated nature of task label generation and diverse subject matter, biases may be minimal but could still exist based on the underlying training data of the zero-shot classifier.", "### Other Known Limitations\nThe primary limitation is the reliance on the zero-shot classifier's accuracy for task labeling, which may not always align with human judgment.", "## Additional Information", "### Dataset Curators\nThe dataset was curated by the team involved in the development of 'mmlu'.", "### Licensing Information\nThe dataset is available under the Apache-2.0 License.\n\n\n@misc{mmlu_auxiliary_trained_set,\n title = {{MMLU Auxiliary Trained Set Labelled by e5-mistral-7b-instruct}},\n author = {Kaizhao Liang},\n year = {2024},\n howpublished = {URL\n note = {Accessed: Date of Access},\n description = {A dataset of 99,842 examples across various subjects, each including a question, multiple choice options, a subject category, an answer, and a task label generated by a zero-shot classifier constructed from the intfloat/e5-mistral-7b-instruct model.},\n license = {Apache-2.0}\n}", "### Contact Information\nhomepage", "### Visualizations\n- Counts by Category\n <img src=\"URL\n\n- Counts by Super Category\n <img src=\"URL\n\n---" ]

[ "license:apache-2.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-apache-2.0 #region-us \n" ]

[ "task_categories:text-classification", "size_categories:1M<n<10M", "language:en", "language:fr", "license:apache-2.0", "Routing", "router", "region:us" ]

[]

[ "en", "fr" ]

[ "### Why this dataset ?\nTraining a multimodal router holds crucial significance in the realm of artificial intelligence. By harmonizing different specialized models within a constellation, the router plays a central role in intelligently orchestrating tasks. This approach not only enables precise classification but also paves the way for diverse applications of artificial intelligence, thereby enhancing our ability to successfully navigate through the complexities of multimodal data. In essence, training a multimodal router represents a vital strategic advancement, unveiling exciting new prospects for the future of AI.\n(Thks ChatGPT)\nMore informations here : URL", "### From :\n- speechcolab/gigaspeech\n- stingning/ultrachat\n- Falah/image_generation_prompts_SDXL", "### Structure \nText\tLabel\n\"Polaroid photo of a middle-aged woman with fa...\t0", "##### 3 Categories : \n0: Image Generation\n1: LLM Generation\n2: Audio Generation" ]

[ "TAGS\n#task_categories-text-classification #size_categories-1M<n<10M #language-English #language-French #license-apache-2.0 #Routing #router #region-us \n", "### Why this dataset ?\nTraining a multimodal router holds crucial significance in the realm of artificial intelligence. By harmonizing different specialized models within a constellation, the router plays a central role in intelligently orchestrating tasks. This approach not only enables precise classification but also paves the way for diverse applications of artificial intelligence, thereby enhancing our ability to successfully navigate through the complexities of multimodal data. In essence, training a multimodal router represents a vital strategic advancement, unveiling exciting new prospects for the future of AI.\n(Thks ChatGPT)\nMore informations here : URL", "### From :\n- speechcolab/gigaspeech\n- stingning/ultrachat\n- Falah/image_generation_prompts_SDXL", "### Structure \nText\tLabel\n\"Polaroid photo of a middle-aged woman with fa...\t0", "##### 3 Categories : \n0: Image Generation\n1: LLM Generation\n2: Audio Generation" ]

[ "task_categories:text-generation", "region:us" ]

[]

[]

[]

[ "TAGS\n#task_categories-text-generation #region-us \n" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"cowese_abrev_binary_topics_LDA_2\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"cowese_abrev_binary_topics_LDA_2\"\n\nMore Information needed" ]

[ "task_categories:text-generation", "size_categories:1K<n<10K", "language:en", "license:mit", "philosophy", "region:us" ]

[]

[ "en" ]

[ "# Plato: philosophy essays from URL\n\nPlato is a corpus of 2.4k high quality philosophy essays from URL." ]

[ "TAGS\n#task_categories-text-generation #size_categories-1K<n<10K #language-English #license-mit #philosophy #region-us \n", "# Plato: philosophy essays from URL\n\nPlato is a corpus of 2.4k high quality philosophy essays from URL." ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"fashion_image_caption-100-v2\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"fashion_image_caption-100-v2\"\n\nMore Information needed" ]

[ "language:en", "license:cc-by-nc-4.0", "genomics", "phage-prediction", "bioinformatics", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for Phage Prediction Dataset", "## Dataset Description\n\nTo train and assess our prediction models, we assembled a comprehensive phage sequence database from diverse sources.\nAs of July 9, 2023, we procured viral sequences and annotations from the RefSeq database. By isolating entries labeled 'phage', we obtained 6,075 contigs. \nOur database was further enriched with the inclusion of the TemPhD database, adding another 192,326 phage contigs extracted from 148,229 assemblies.\n\nTo address sequence redundancy present in both the RefSeq and TemPhD databases, we applied the CD-HIT algorithm (using CD-HIT-EST with a default word size of 5). \nWhile several clustering thresholds (0.99, 0.95, 0.90) were experimented with and found to produce similar outcomes, we settled on a threshold of 0.99. \nThis process resulted in a refined set of 40,512 distinct phage sequences, with an average length of approximately 43,356 base pairs, culminating in a total of 3.5 billion base pairs.\nNotably, these sequences target a wide spectrum of 660 bacterial genera. Subsequent to sequence curation, phage sequences were mapped to their respective bacterial hosts.\n\nIt is a sample dataset consisting of 10,000 segments, representing random samples and segments of phage genomes.\nThe full dataset is available on Zenodo.", "## Features\n- Phage-Host Associations: Our dataset represents bacteriophages and their bacterial hosts.\n\n- Balanced Representation: The dataset is structured to mitigate bias by evenly representing phages and their hosts across various genera, incorporating reverse-complement sequences for completeness.\n\n- Dataset Composition: The final collection includes sequences of varying lengths to accommodate different research needs, with a balanced distribution across training, validation, and testing sets.\n\n- Sampling Strategy: To ensure a comprehensive yet manageable dataset, we performed undersampling and upsampling techniques, creating a diverse array of sequence lengths and ensuring no overlap between training and testing sets at the species level.", "### Dataset Structure\n\nThe dataset is divided into three subsets based on segment lengths: 512, 1024, and 2048 base pairs. These subsets are named 'sample_test_L512', 'sample_test_L1024', and 'sample_test_L2048', respectively.", "#### Data Fields\n\n- 'segment_id': Unique identifier for each genomic segment.\n- 'contig_id': Identifier for the contig from which the segment is derived.\n- 'segment_start': Start position of the segment in the contig.\n- 'segment_end': End position of the segment in the contig.\n- 'L': Length of the genomic segment (512, 1024, or 2048).\n- 'segment': The genomic sequence of the segment.\n- 'label': Classification label (e.g., 'phage').\n- 'y': Binary label (1 for phage, 0 for non-phage).", "### Data Splits\n\nThe dataset is structured as follows:\n\n- 'sample_test_L512': Test set with segment length of 512.\n- 'sample_test_L1024': Test set with segment length of 1024.\n- 'sample_test_L2048': Test set with segment length of 2048.", "## Dataset Creation", "### Source Data\n\nThe dataset is compiled from diverse genomic sources, with a focus on phage sequences and annotations from the RefSeq database and a dataset validated through the TemPhD method. Redundancy in sequences is addressed using the CD-HIT algorithm.", "## Contact Information\n\nFor any questions, feedback, or contributions regarding the datasets or ProkBERT, please feel free to reach out:\n\n- Name: Balázs Ligeti\n- Email: obalasz@URL\n\nWe welcome your input and collaboration to improve our resources and research.\n\n\n\n'''bibtex\n@Article{ProkBERT2024,\n author = {Ligeti, Balázs et al.},\n journal = {Frontiers in Microbiology},\n title = {{ProkBERT} family: genomic language models},\n year = {2024},\n volume = {14},\n URL = {URL\n DOI = {10.3389/fmicb.2023.1331233}\n}" ]

[ "TAGS\n#language-English #license-cc-by-nc-4.0 #genomics #phage-prediction #bioinformatics #region-us \n", "# Dataset Card for Phage Prediction Dataset", "## Dataset Description\n\nTo train and assess our prediction models, we assembled a comprehensive phage sequence database from diverse sources.\nAs of July 9, 2023, we procured viral sequences and annotations from the RefSeq database. By isolating entries labeled 'phage', we obtained 6,075 contigs. \nOur database was further enriched with the inclusion of the TemPhD database, adding another 192,326 phage contigs extracted from 148,229 assemblies.\n\nTo address sequence redundancy present in both the RefSeq and TemPhD databases, we applied the CD-HIT algorithm (using CD-HIT-EST with a default word size of 5). \nWhile several clustering thresholds (0.99, 0.95, 0.90) were experimented with and found to produce similar outcomes, we settled on a threshold of 0.99. \nThis process resulted in a refined set of 40,512 distinct phage sequences, with an average length of approximately 43,356 base pairs, culminating in a total of 3.5 billion base pairs.\nNotably, these sequences target a wide spectrum of 660 bacterial genera. Subsequent to sequence curation, phage sequences were mapped to their respective bacterial hosts.\n\nIt is a sample dataset consisting of 10,000 segments, representing random samples and segments of phage genomes.\nThe full dataset is available on Zenodo.", "## Features\n- Phage-Host Associations: Our dataset represents bacteriophages and their bacterial hosts.\n\n- Balanced Representation: The dataset is structured to mitigate bias by evenly representing phages and their hosts across various genera, incorporating reverse-complement sequences for completeness.\n\n- Dataset Composition: The final collection includes sequences of varying lengths to accommodate different research needs, with a balanced distribution across training, validation, and testing sets.\n\n- Sampling Strategy: To ensure a comprehensive yet manageable dataset, we performed undersampling and upsampling techniques, creating a diverse array of sequence lengths and ensuring no overlap between training and testing sets at the species level.", "### Dataset Structure\n\nThe dataset is divided into three subsets based on segment lengths: 512, 1024, and 2048 base pairs. These subsets are named 'sample_test_L512', 'sample_test_L1024', and 'sample_test_L2048', respectively.", "#### Data Fields\n\n- 'segment_id': Unique identifier for each genomic segment.\n- 'contig_id': Identifier for the contig from which the segment is derived.\n- 'segment_start': Start position of the segment in the contig.\n- 'segment_end': End position of the segment in the contig.\n- 'L': Length of the genomic segment (512, 1024, or 2048).\n- 'segment': The genomic sequence of the segment.\n- 'label': Classification label (e.g., 'phage').\n- 'y': Binary label (1 for phage, 0 for non-phage).", "### Data Splits\n\nThe dataset is structured as follows:\n\n- 'sample_test_L512': Test set with segment length of 512.\n- 'sample_test_L1024': Test set with segment length of 1024.\n- 'sample_test_L2048': Test set with segment length of 2048.", "## Dataset Creation", "### Source Data\n\nThe dataset is compiled from diverse genomic sources, with a focus on phage sequences and annotations from the RefSeq database and a dataset validated through the TemPhD method. Redundancy in sequences is addressed using the CD-HIT algorithm.", "## Contact Information\n\nFor any questions, feedback, or contributions regarding the datasets or ProkBERT, please feel free to reach out:\n\n- Name: Balázs Ligeti\n- Email: obalasz@URL\n\nWe welcome your input and collaboration to improve our resources and research.\n\n\n\n'''bibtex\n@Article{ProkBERT2024,\n author = {Ligeti, Balázs et al.},\n journal = {Frontiers in Microbiology},\n title = {{ProkBERT} family: genomic language models},\n year = {2024},\n volume = {14},\n URL = {URL\n DOI = {10.3389/fmicb.2023.1331233}\n}" ]

[ "license:mit", "region:us" ]

[]

[]

[ "# Quantum Electronic Integrals\n\nThe files contained in this dataset \nwere produced with URL.", "## Mono-Electronic\n\nSee mono for 1-electron integrals. \nEach JSON file consists of a list of objects of the following form" ]

[ "TAGS\n#license-mit #region-us \n", "# Quantum Electronic Integrals\n\nThe files contained in this dataset \nwere produced with URL.", "## Mono-Electronic\n\nSee mono for 1-electron integrals. \nEach JSON file consists of a list of objects of the following form" ]

[ "license:cc-by-nc-4.0", "genomics", "ESKAPE pathogens", "bioinformatics", "ProkBERT", "region:us" ]

[]

[]

[ "# Dataset Card for ESKAPE Genomic Features Dataset", "## Dataset Description\n\nThis dataset includes genomic segments from ESKAPE pathogens, characterized by various genomic features such as coding sequences (CDS), intergenic regions, ncRNA, and pseudogenes. It was analyzed to understand the representations captured by models like ProkBERT-mini, ProkBERT-mini-c, and ProkBERT-mini-long.", "### Data Fields\n\n- 'contig_id': Identifier of the contig.\n- 'segment_id': Unique identifier for each genomic segment.\n- 'strand': DNA strand of the segment ('+' or '-').\n- 'seq_start': Starting position of the segment in the contig.\n- 'seq_end': Ending position of the segment in the contig.\n- 'segment_start': Starting position of the segment in the sequence.\n- 'segment_end': Ending position of the segment in the sequence.\n- 'label': Genomic feature category (e.g., CDS, intergenic).\n- 'segment_length': Length of the genomic segment.\n- 'Nsegment': Length of the genomic segment.\n- 'segment': Genomic sequence of the segment.", "### UMAP Embeddings and Silhouette Scores\n\nThe dataset was used to assess the zero-shot capabilities of the ProkBERT models in predicting genomic features. UMAP technique was employed to reduce dimensionality and derive embeddings, which were then evaluated using silhouette scores. The embeddings and scores reveal the models' proficiency in differentiating between genomic features and capturing the genomic structure of ESKAPE pathogens.", "## Dataset Creation\n\nThe dataset is compiled from the RefSeq database and other sources, focusing on ESKAPE pathogens. The genomic features were sampled randomly, followed by contigous segmentation. The segment length is 256, shorter fragments were discarded.", "## Overview of ESKAPE Pathogens\n\nESKAPE pathogens are a group of bacteria that pose a significant threat to public health due to their high levels of antibiotic resistance. The acronym ESKAPE represents six genera of bacteria:\n\n- Enterococcus faecium\n- Staphylococcus aureus\n- Klebsiella pneumoniae\n- Acinetobacter baumannii\n- Pseudomonas aeruginosa\n- Enterobacter species\n\nThese pathogens are known for \"escaping\" the effects of antibiotics and are responsible for a large proportion of nosocomial infections (hospital-acquired infections). They are particularly concerning in healthcare settings because they can lead to severe infections that are increasingly difficult to treat due to their resistance to multiple antibiotics.", "## Considerations for Using the Data\n\nThis dataset is relevant for genomic research and bioinformatics, particularly for understanding the genomic structure of ESKAPE pathogens and their representation in embedding spaces.", "## Contact Information\n\nFor inquiries or feedback regarding this dataset, please contact:\n\n- Balázs Ligeti\n- Email: obalasz@URL", "### Dataset Curators\n\nThis dataset was curated by Balázs Ligeti from the Neural Bioinformatics Research Group, Faculty of Information Technology and Bionics, Pázmány Péter Catholic University (PPCU-FITB).\n\n\n\n\n\nIf you use the code or data in this package, please cite:" ]

[ "TAGS\n#license-cc-by-nc-4.0 #genomics #ESKAPE pathogens #bioinformatics #ProkBERT #region-us \n", "# Dataset Card for ESKAPE Genomic Features Dataset", "## Dataset Description\n\nThis dataset includes genomic segments from ESKAPE pathogens, characterized by various genomic features such as coding sequences (CDS), intergenic regions, ncRNA, and pseudogenes. It was analyzed to understand the representations captured by models like ProkBERT-mini, ProkBERT-mini-c, and ProkBERT-mini-long.", "### Data Fields\n\n- 'contig_id': Identifier of the contig.\n- 'segment_id': Unique identifier for each genomic segment.\n- 'strand': DNA strand of the segment ('+' or '-').\n- 'seq_start': Starting position of the segment in the contig.\n- 'seq_end': Ending position of the segment in the contig.\n- 'segment_start': Starting position of the segment in the sequence.\n- 'segment_end': Ending position of the segment in the sequence.\n- 'label': Genomic feature category (e.g., CDS, intergenic).\n- 'segment_length': Length of the genomic segment.\n- 'Nsegment': Length of the genomic segment.\n- 'segment': Genomic sequence of the segment.", "### UMAP Embeddings and Silhouette Scores\n\nThe dataset was used to assess the zero-shot capabilities of the ProkBERT models in predicting genomic features. UMAP technique was employed to reduce dimensionality and derive embeddings, which were then evaluated using silhouette scores. The embeddings and scores reveal the models' proficiency in differentiating between genomic features and capturing the genomic structure of ESKAPE pathogens.", "## Dataset Creation\n\nThe dataset is compiled from the RefSeq database and other sources, focusing on ESKAPE pathogens. The genomic features were sampled randomly, followed by contigous segmentation. The segment length is 256, shorter fragments were discarded.", "## Overview of ESKAPE Pathogens\n\nESKAPE pathogens are a group of bacteria that pose a significant threat to public health due to their high levels of antibiotic resistance. The acronym ESKAPE represents six genera of bacteria:\n\n- Enterococcus faecium\n- Staphylococcus aureus\n- Klebsiella pneumoniae\n- Acinetobacter baumannii\n- Pseudomonas aeruginosa\n- Enterobacter species\n\nThese pathogens are known for \"escaping\" the effects of antibiotics and are responsible for a large proportion of nosocomial infections (hospital-acquired infections). They are particularly concerning in healthcare settings because they can lead to severe infections that are increasingly difficult to treat due to their resistance to multiple antibiotics.", "## Considerations for Using the Data\n\nThis dataset is relevant for genomic research and bioinformatics, particularly for understanding the genomic structure of ESKAPE pathogens and their representation in embedding spaces.", "## Contact Information\n\nFor inquiries or feedback regarding this dataset, please contact:\n\n- Balázs Ligeti\n- Email: obalasz@URL", "### Dataset Curators\n\nThis dataset was curated by Balázs Ligeti from the Neural Bioinformatics Research Group, Faculty of Information Technology and Bionics, Pázmány Péter Catholic University (PPCU-FITB).\n\n\n\n\n\nIf you use the code or data in this package, please cite:" ]

[ "language:en", "license:unknown", "legal", "region:us" ]

[]

[ "en" ]

[ "# North Carolina Police Incident Dataset", "## Dataset Description\n\n- Homepage: The processed dataset, which will soon be updated, is available at the following Huggingface portal: URL For the raw datasets, their respective homepages are:\n - Cary:\n - Cary Open Data Portal - CPD Incidents\n - Chapel Hill:\n - Chapel Hill Open Data Portal - Police Incidents\n - Charlotte:\n - Charlotte Open Data Portal - CMPD Incidents\n - Durham:\n - Durham Open Data Portal - DPD Incidents UCR/NIBRS Reporting\n - Raleigh:\n - Raleigh Open Data Portal - Police Incidents (NIBRS)\n \n- Point of Contact: For any issues related to the raw datasets, please reach out to the respective government offices. For inquiries or issues regarding the processed data, you can contact zwn22 at Huggingface.", "### Dataset Summary\n\nThe dataset, derived from public police incident reports from various cities in North Carolina, spans from the early 2000s to 2024. The source data originate from the open data portals of Cary, Chapel Hill, Charlotte, Durham, and Raleigh. The dataset underwent a process of merging data from these cities, followed by cleaning to remove incomplete rows. Additionally, the dataset focuses on extracting and categorizing major crime types, providing valuable information such as crime type, time, location of occurrence, and other relevant details.", "### Supported Tasks\n\n1. Crime Trend Analysis: Analyzing crime trends over time and across different locations. This could involve identifying patterns in crime rates, seasonal variations, or shifts in the types of crimes committed.\n\n2. Predictive Policing: Developing models to predict future crime occurrences based on historical data. This could help in resource allocation and proactive policing strategies.\n\n3. Geospatial Analysis: Mapping crime incidents to identify hotspots and regions with higher crime rates. This can aid in understanding geographical factors influencing crime and in deploying resources more effectively.", "### Languages\n\nEnglish", "## Dataset Structure", "### Data Instances\n\nHere is an illustrative example from the processed dataset (note that specific details are subject to change):", "### Data Fields\n\nThe dataset contains several fields, each providing specific information about police incidents. Here is a list of these fields along with their descriptions and data types:\n\n- 'year' (integer): The year in which the incident occurred. Used as input in temporal analysis tasks.\n- 'city' (string): The city where the incident took place. This field is crucial for geographic analyses and comparisons between cities.\n- 'crime_major_category' (string): A broad categorization of the crime, used as input for crime pattern analysis and categorization tasks.\n- 'crime_specific_category' (string): More detailed classification of the crime, falling under the major category. This field allows for a finer-grained analysis of crime types.\n- 'latitude' (float) and 'longitude' (float): Geographical coordinates pinpointing the location of the incident. These fields are essential for geospatial analysis.\n- 'start_time' (datetime) and 'end_time' (datetime): The beginning and end times of the incident, providing temporal context. These fields are used in analyses that require time-based information.\n- 'clear_status' (string): The resolution status of the case, such as whether it was cleared by arrest or remains under investigation. This field can be used to understand case outcomes.\n- 'incident_address' (string): The specific address where the incident occurred. This field adds a detailed spatial dimension to the data.\n- 'notes' (string): Additional remarks or details about the incident, like weapon usage or other relevant factors. This field provides supplementary information that may be relevant for certain analyses.", "## Dataset Creation", "### Curation Rationale\n\nThe dataset, covering police incidents in select North Carolina cities from 2000 to 2024, aims to aid crime research. It provides a long-term view of crime patterns and trends, useful for criminologists, sociologists, and public policy researchers. The comprehensive data enables analyses of crime evolution and its socio-economic correlations. It also supports the development of predictive models for law enforcement and policy planning. Additionally, the dataset's multi-city scope allows for comparative studies to understand unique challenges and inform localized crime prevention strategies.", "### Source Data\n\nFive datasets are primarily utilized as source data:\n- Cary:\n - Cary Open Data Portal - CPD Incidents\n - Details:\n - Size: 116317 rows * 34 columns\n - Column names: 'Crime Category', 'Crime Type', 'UCR', 'Map Reference',\n 'Incident Number', 'Begin Date Of Occurrence',\n 'Begin Time Of Occurrence', 'End Date Of Occurrence',\n 'End Time Of Occurrence', 'Crime Day', 'Geo Code', 'Location Category',\n 'District', 'Beat Number', 'Location', 'ID', 'Lat', 'Lon',\n 'Charge Count', 'Neighborhood ID', 'Apartment Complex',\n 'Residential Subdivision', 'Subdivision ID', 'Phx Activity Date',\n 'Phx Record Status', 'Phx Community', 'Phx Status', 'Record',\n 'Offense Category', 'Violent Property', 'timeframe', 'domestic',\n 'Total Incidents', 'Year'\n- Chapel Hill:\n - Chapel Hill Open Data Portal - Police Incidents\n - Details:\n - Size: 101828 rows * 19 columns\n - Column names: 'Incident ID', 'Agency', 'Offense', 'Street', 'City', 'State', 'Zipcode', 'Date of Report', 'Date of Occurrence', 'Date Found', 'Reported As', 'Premise Description', 'Forcible', 'Weapon Description', 'Victim Age', 'Victim Race', 'Victim Gender', 'Latitude', 'Longitude'\n- Charlotte:\n - Charlotte Open Data Portal - CMPD Incidents\n - Details:\n - Size: 483632 rows * 30 columns\n - Column names: 'X', 'Y', 'YEAR', 'INCIDENT_REPORT_ID', 'LOCATION', 'CITY', 'STATE',\n 'ZIP', 'X_COORD_PUBLIC', 'Y_COORD_PUBLIC', 'LATITUDE_PUBLIC',\n 'LONGITUDE_PUBLIC', 'DIVISION_ID', 'CMPD_PATROL_DIVISION', 'NPA',\n 'DATE_REPORTED', 'DATE_INCIDENT_BEGAN', 'DATE_INCIDENT_END',\n 'ADDRESS_DESCRIPTION', 'LOCATION_TYPE_DESCRIPTION',\n 'PLACE_TYPE_DESCRIPTION', 'PLACE_DETAIL_DESCRIPTION',\n 'CLEARANCE_STATUS', 'CLEARANCE_DETAIL_STATUS', 'CLEARANCE_DATE',\n 'HIGHEST_NIBRS_CODE', 'HIGHEST_NIBRS_DESCRIPTION', 'OBJECTID', 'Shape',\n 'GlobalID'\n- Durham:\n - Durham Open Data Portal - DPD Incidents UCR/NIBRS Reporting\n - Details:\n - Size: 149924 rows * 16 columns\n - Column names: 'Case Number', 'Report Date', 'Report Time', 'Status', 'Sequence',\n 'ATT/COM', 'UCR Code', 'Offense', 'Address', 'X', 'Y', 'District',\n 'Beat', 'Tract', 'Premise', 'Weapon'\n- Raleigh:\n - Raleigh Open Data Portal - Police Incidents (NIBRS)\n - Details:\n - Size: 493912 rows * 19 columns\n - Column names: 'Case Number', 'Crime_Category', 'Crime Code', 'Crime Description', 'Crime Type', 'Reported Block Address', 'City of Incident', 'City', 'District', 'Reported Date', 'Reported Year', 'Reported Month', 'Reported Day', 'Reported Hour', 'Reported Day of Week', 'Latitude', 'Longitude', 'Agency', 'Updated_Date'", "## Considerations for Using the Data", "### Other Known Limitations\n\nThe interpretation rights of the dataset are reserved by the respective government authorities. It is subject to change, and the City of Raleigh, as an example, retains the right to modify or discontinue any of the data feeds at any given time. This includes the right to require termination of displaying, distributing, or using the data, for any reason, including but not limited to violations of the Terms of Use. Users should be aware that such changes can occur and that the dataset may evolve over time according to the decisions made by the governing bodies.", "## Additional Information", "### Dataset Curators\n\nFor detailed information regarding the individuals involved in collecting the dataset and their affiliations, as well as any funding details, interested parties are encouraged to directly contact the respective government offices.", "### Licensing Information\n\nThe dataset from cities in North Carolina, including Cary, Charlotte, Durham, and Raleigh, is sourced from their respective open data portals, each with its specific licensing information and terms of use.\n\nCary's Open Data Portal operates under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license. This allows for both commercial and non-commercial use, distribution, and adaptation of the data, as long as proper attribution is given and any derived works are shared under the same terms.\n\nChapel Hill's Open Data Portal is governed by the Open Database License (ODbL) 1.0 License. This license permits both commercial and non-commercial use, distribution, and modification of the data. Users are required to attribute the data appropriately and, if the data is altered or transformed, or if new datasets are built upon it, the resulting work must be distributed under the same ODbL 1.0 License. \n\nWhile Durham and Charlotte's Open Data Portals do not specify a particular license, their websites state that the data is publicly accessible and can be freely used. Charlotte's portal serves as a resource for the public to explore, visualize, and download data, facilitating the creation of new applications and maps. Similarly, Durham's portal is dedicated to making all government data available for public use without restrictions.\n\nRaleigh's Open Data Policy stipulates that by using data from their site, users agree to the terms and conditions described under the City of Raleigh Open Data Privacy Policy. The City of Raleigh makes no warranties regarding the completeness, accuracy, or timeliness of the data, and disclaims all express or implied warranties, including those of merchantability or fitness for a particular purpose. Users should be aware that the data format or schema may change as updates are made, and they use the information at their own risk.", "### Contributions\n\nThanks to Town of Cary, City of Chapel Hill, City of Charlotte, City of Durham, and City of Raleigh for providing the raw dataset." ]

[ "TAGS\n#language-English #license-unknown #legal #region-us \n", "# North Carolina Police Incident Dataset", "## Dataset Description\n\n- Homepage: The processed dataset, which will soon be updated, is available at the following Huggingface portal: URL For the raw datasets, their respective homepages are:\n - Cary:\n - Cary Open Data Portal - CPD Incidents\n - Chapel Hill:\n - Chapel Hill Open Data Portal - Police Incidents\n - Charlotte:\n - Charlotte Open Data Portal - CMPD Incidents\n - Durham:\n - Durham Open Data Portal - DPD Incidents UCR/NIBRS Reporting\n - Raleigh:\n - Raleigh Open Data Portal - Police Incidents (NIBRS)\n \n- Point of Contact: For any issues related to the raw datasets, please reach out to the respective government offices. For inquiries or issues regarding the processed data, you can contact zwn22 at Huggingface.", "### Dataset Summary\n\nThe dataset, derived from public police incident reports from various cities in North Carolina, spans from the early 2000s to 2024. The source data originate from the open data portals of Cary, Chapel Hill, Charlotte, Durham, and Raleigh. The dataset underwent a process of merging data from these cities, followed by cleaning to remove incomplete rows. Additionally, the dataset focuses on extracting and categorizing major crime types, providing valuable information such as crime type, time, location of occurrence, and other relevant details.", "### Supported Tasks\n\n1. Crime Trend Analysis: Analyzing crime trends over time and across different locations. This could involve identifying patterns in crime rates, seasonal variations, or shifts in the types of crimes committed.\n\n2. Predictive Policing: Developing models to predict future crime occurrences based on historical data. This could help in resource allocation and proactive policing strategies.\n\n3. Geospatial Analysis: Mapping crime incidents to identify hotspots and regions with higher crime rates. This can aid in understanding geographical factors influencing crime and in deploying resources more effectively.", "### Languages\n\nEnglish", "## Dataset Structure", "### Data Instances\n\nHere is an illustrative example from the processed dataset (note that specific details are subject to change):", "### Data Fields\n\nThe dataset contains several fields, each providing specific information about police incidents. Here is a list of these fields along with their descriptions and data types:\n\n- 'year' (integer): The year in which the incident occurred. Used as input in temporal analysis tasks.\n- 'city' (string): The city where the incident took place. This field is crucial for geographic analyses and comparisons between cities.\n- 'crime_major_category' (string): A broad categorization of the crime, used as input for crime pattern analysis and categorization tasks.\n- 'crime_specific_category' (string): More detailed classification of the crime, falling under the major category. This field allows for a finer-grained analysis of crime types.\n- 'latitude' (float) and 'longitude' (float): Geographical coordinates pinpointing the location of the incident. These fields are essential for geospatial analysis.\n- 'start_time' (datetime) and 'end_time' (datetime): The beginning and end times of the incident, providing temporal context. These fields are used in analyses that require time-based information.\n- 'clear_status' (string): The resolution status of the case, such as whether it was cleared by arrest or remains under investigation. This field can be used to understand case outcomes.\n- 'incident_address' (string): The specific address where the incident occurred. This field adds a detailed spatial dimension to the data.\n- 'notes' (string): Additional remarks or details about the incident, like weapon usage or other relevant factors. This field provides supplementary information that may be relevant for certain analyses.", "## Dataset Creation", "### Curation Rationale\n\nThe dataset, covering police incidents in select North Carolina cities from 2000 to 2024, aims to aid crime research. It provides a long-term view of crime patterns and trends, useful for criminologists, sociologists, and public policy researchers. The comprehensive data enables analyses of crime evolution and its socio-economic correlations. It also supports the development of predictive models for law enforcement and policy planning. Additionally, the dataset's multi-city scope allows for comparative studies to understand unique challenges and inform localized crime prevention strategies.", "### Source Data\n\nFive datasets are primarily utilized as source data:\n- Cary:\n - Cary Open Data Portal - CPD Incidents\n - Details:\n - Size: 116317 rows * 34 columns\n - Column names: 'Crime Category', 'Crime Type', 'UCR', 'Map Reference',\n 'Incident Number', 'Begin Date Of Occurrence',\n 'Begin Time Of Occurrence', 'End Date Of Occurrence',\n 'End Time Of Occurrence', 'Crime Day', 'Geo Code', 'Location Category',\n 'District', 'Beat Number', 'Location', 'ID', 'Lat', 'Lon',\n 'Charge Count', 'Neighborhood ID', 'Apartment Complex',\n 'Residential Subdivision', 'Subdivision ID', 'Phx Activity Date',\n 'Phx Record Status', 'Phx Community', 'Phx Status', 'Record',\n 'Offense Category', 'Violent Property', 'timeframe', 'domestic',\n 'Total Incidents', 'Year'\n- Chapel Hill:\n - Chapel Hill Open Data Portal - Police Incidents\n - Details:\n - Size: 101828 rows * 19 columns\n - Column names: 'Incident ID', 'Agency', 'Offense', 'Street', 'City', 'State', 'Zipcode', 'Date of Report', 'Date of Occurrence', 'Date Found', 'Reported As', 'Premise Description', 'Forcible', 'Weapon Description', 'Victim Age', 'Victim Race', 'Victim Gender', 'Latitude', 'Longitude'\n- Charlotte:\n - Charlotte Open Data Portal - CMPD Incidents\n - Details:\n - Size: 483632 rows * 30 columns\n - Column names: 'X', 'Y', 'YEAR', 'INCIDENT_REPORT_ID', 'LOCATION', 'CITY', 'STATE',\n 'ZIP', 'X_COORD_PUBLIC', 'Y_COORD_PUBLIC', 'LATITUDE_PUBLIC',\n 'LONGITUDE_PUBLIC', 'DIVISION_ID', 'CMPD_PATROL_DIVISION', 'NPA',\n 'DATE_REPORTED', 'DATE_INCIDENT_BEGAN', 'DATE_INCIDENT_END',\n 'ADDRESS_DESCRIPTION', 'LOCATION_TYPE_DESCRIPTION',\n 'PLACE_TYPE_DESCRIPTION', 'PLACE_DETAIL_DESCRIPTION',\n 'CLEARANCE_STATUS', 'CLEARANCE_DETAIL_STATUS', 'CLEARANCE_DATE',\n 'HIGHEST_NIBRS_CODE', 'HIGHEST_NIBRS_DESCRIPTION', 'OBJECTID', 'Shape',\n 'GlobalID'\n- Durham:\n - Durham Open Data Portal - DPD Incidents UCR/NIBRS Reporting\n - Details:\n - Size: 149924 rows * 16 columns\n - Column names: 'Case Number', 'Report Date', 'Report Time', 'Status', 'Sequence',\n 'ATT/COM', 'UCR Code', 'Offense', 'Address', 'X', 'Y', 'District',\n 'Beat', 'Tract', 'Premise', 'Weapon'\n- Raleigh:\n - Raleigh Open Data Portal - Police Incidents (NIBRS)\n - Details:\n - Size: 493912 rows * 19 columns\n - Column names: 'Case Number', 'Crime_Category', 'Crime Code', 'Crime Description', 'Crime Type', 'Reported Block Address', 'City of Incident', 'City', 'District', 'Reported Date', 'Reported Year', 'Reported Month', 'Reported Day', 'Reported Hour', 'Reported Day of Week', 'Latitude', 'Longitude', 'Agency', 'Updated_Date'", "## Considerations for Using the Data", "### Other Known Limitations\n\nThe interpretation rights of the dataset are reserved by the respective government authorities. It is subject to change, and the City of Raleigh, as an example, retains the right to modify or discontinue any of the data feeds at any given time. This includes the right to require termination of displaying, distributing, or using the data, for any reason, including but not limited to violations of the Terms of Use. Users should be aware that such changes can occur and that the dataset may evolve over time according to the decisions made by the governing bodies.", "## Additional Information", "### Dataset Curators\n\nFor detailed information regarding the individuals involved in collecting the dataset and their affiliations, as well as any funding details, interested parties are encouraged to directly contact the respective government offices.", "### Licensing Information\n\nThe dataset from cities in North Carolina, including Cary, Charlotte, Durham, and Raleigh, is sourced from their respective open data portals, each with its specific licensing information and terms of use.\n\nCary's Open Data Portal operates under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license. This allows for both commercial and non-commercial use, distribution, and adaptation of the data, as long as proper attribution is given and any derived works are shared under the same terms.\n\nChapel Hill's Open Data Portal is governed by the Open Database License (ODbL) 1.0 License. This license permits both commercial and non-commercial use, distribution, and modification of the data. Users are required to attribute the data appropriately and, if the data is altered or transformed, or if new datasets are built upon it, the resulting work must be distributed under the same ODbL 1.0 License. \n\nWhile Durham and Charlotte's Open Data Portals do not specify a particular license, their websites state that the data is publicly accessible and can be freely used. Charlotte's portal serves as a resource for the public to explore, visualize, and download data, facilitating the creation of new applications and maps. Similarly, Durham's portal is dedicated to making all government data available for public use without restrictions.\n\nRaleigh's Open Data Policy stipulates that by using data from their site, users agree to the terms and conditions described under the City of Raleigh Open Data Privacy Policy. The City of Raleigh makes no warranties regarding the completeness, accuracy, or timeliness of the data, and disclaims all express or implied warranties, including those of merchantability or fitness for a particular purpose. Users should be aware that the data format or schema may change as updates are made, and they use the information at their own risk.", "### Contributions\n\nThanks to Town of Cary, City of Chapel Hill, City of Charlotte, City of Durham, and City of Raleigh for providing the raw dataset." ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "license:mit", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-mit #region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:text-generation", "language:de", "license:mit", "region:us" ]

[]

[ "de" ]

[]

[ "TAGS\n#task_categories-text-generation #language-German #license-mit #region-us \n" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"Baarat-Hin-QA\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"Baarat-Hin-QA\"\n\nMore Information needed" ]

[ "task_categories:text-to-image", "size_categories:n<1K", "license:cc-by-4.0", "art", "not-for-all-audiences", "region:us" ]

[]

[]

[ "### Disclaimer\n\n\n* This dataset is intented to be used in generative AI - text-to-image models, it was created with the intended purpose of making a Stable-diffusion LoRA model.\n* the dataset was built upon the Dreambooth caption method, the dataset follows this structure:\n\n\nLicense\n=======\n\n\nThis dataset is provided under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license." ]

[ "TAGS\n#task_categories-text-to-image #size_categories-n<1K #license-cc-by-4.0 #art #not-for-all-audiences #region-us \n", "### Disclaimer\n\n\n* This dataset is intented to be used in generative AI - text-to-image models, it was created with the intended purpose of making a Stable-diffusion LoRA model.\n* the dataset was built upon the Dreambooth caption method, the dataset follows this structure:\n\n\nLicense\n=======\n\n\nThis dataset is provided under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license." ]

[ "Lilac", "region:us" ]

[]

[]

[ "# local/mikeion_dissertation_data\nThis dataset is a Lilac processed dataset. Original dataset: URL\n\nTo download the dataset to a local directory:\n\n\n\nor from python with:" ]

[ "TAGS\n#Lilac #region-us \n", "# local/mikeion_dissertation_data\nThis dataset is a Lilac processed dataset. Original dataset: URL\n\nTo download the dataset to a local directory:\n\n\n\nor from python with:" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:question-answering", "language:en", "license:mit", "motorcycles", "repair", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#task_categories-question-answering #language-English #license-mit #motorcycles #repair #region-us \n" ]

[ "Lilac", "region:us" ]

[]

[]

[ "# lilac/OpenHermes-2.5\nThis dataset is a Lilac processed dataset. Original dataset: URL\n\nTo download the dataset to a local directory:\n\n\n\nor from python with:" ]

[ "TAGS\n#Lilac #region-us \n", "# lilac/OpenHermes-2.5\nThis dataset is a Lilac processed dataset. Original dataset: URL\n\nTo download the dataset to a local directory:\n\n\n\nor from python with:" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"processed_argilla\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"processed_argilla\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"ProteinDataset\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"ProteinDataset\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"FineTuneDatasetProtein\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"FineTuneDatasetProtein\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"CASFPocket\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"CASFPocket\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"CASFProtein\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"CASFProtein\"\n\nMore Information needed" ]

[ "size_categories:n<1K", "region:us" ]

[]

[]

[ "# Dataset Card for Dataset Name\n\n\n\nThis dataset card aims to be a base template for new datasets. It has been generated using this raw template.", "## Dataset Details", "### Dataset Description\n\n\n\n\n\n- Curated by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Language(s) (NLP): \n- License:", "### Dataset Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Out-of-Scope Use", "## Dataset Structure", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Data Collection and Processing", "#### Who are the source data producers?", "### Annotations [optional]", "#### Annotation process", "#### Who are the annotators?", "#### Personal and Sensitive Information", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Dataset Card Authors [optional]", "## Dataset Card Contact" ]

[ "TAGS\n#size_categories-n<1K #region-us \n", "# Dataset Card for Dataset Name\n\n\n\nThis dataset card aims to be a base template for new datasets. It has been generated using this raw template.", "## Dataset Details", "### Dataset Description\n\n\n\n\n\n- Curated by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Language(s) (NLP): \n- License:", "### Dataset Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Out-of-Scope Use", "## Dataset Structure", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Data Collection and Processing", "#### Who are the source data producers?", "### Annotations [optional]", "#### Annotation process", "#### Who are the annotators?", "#### Personal and Sensitive Information", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Dataset Card Authors [optional]", "## Dataset Card Contact" ]

[ "Lilac", "region:us" ]

[]

[]

[ "# local/distilabel-capybara-dpo-7k-binarized\nThis dataset is a Lilac processed dataset. Original dataset: URL\n\nTo download the dataset to a local directory:\n\n\n\nor from python with:" ]

[ "TAGS\n#Lilac #region-us \n", "# local/distilabel-capybara-dpo-7k-binarized\nThis dataset is a Lilac processed dataset. Original dataset: URL\n\nTo download the dataset to a local directory:\n\n\n\nor from python with:" ]

[ "task_categories:other", "annotations_creators:crowdsourced", "annotations_creators:expert-generated", "language_creators:crowdsourced", "language_creators:expert-generated", "multilinguality:multilingual", "size_categories:100K<n<1M", "source_datasets:original", "language:amh", "language:arb", "language:ary", "language:ars", "language:acq", "language:arz", "language:apc", "language:ben", "language:ceb", "language:dan", "language:deu", "language:ell", "language:eng", "language:eus", "language:fil", "language:fin", "language:fra", "language:gle", "language:guj", "language:hat", "language:hau", "language:hin", "language:hun", "language:ibo", "language:ind", "language:ita", "language:jav", "language:jpn", "language:kan", "language:kir", "language:kor", "language:kur", "language:lit", "language:mal", "language:mar", "language:mlg", "language:msa", "language:mya", "language:nep", "language:nld", "language:nso", "language:nya", "language:pan", "language:pes", "language:pol", "language:por", "language:pus", "language:rus", "language:sin", "language:sna", "language:snd", "language:som", "language:spa", "language:sqi", "language:srp", "language:sun", "language:swa", "language:swe", "language:tam", "language:tel", "language:tha", "language:tur", "language:ukr", "language:urd", "language:vie", "language:wol", "language:xho", "language:yor", "language:zho", "language:zul", "license:apache-2.0", "arxiv:2402.06619", "region:us" ]

[ "2402.06619" ]

[ "amh", "arb", "ary", "ars", "acq", "arz", "apc", "ben", "ceb", "dan", "deu", "ell", "eng", "eus", "fil", "fin", "fra", "gle", "guj", "hat", "hau", "hin", "hun", "ibo", "ind", "ita", "jav", "jpn", "kan", "kir", "kor", "kur", "lit", "mal", "mar", "mlg", "msa", "mya", "nep", "nld", "nso", "nya", "pan", "pes", "pol", "por", "pus", "rus", "sin", "sna", "snd", "som", "spa", "sqi", "srp", "sun", "swa", "swe", "tam", "tel", "tha", "tur", "ukr", "urd", "vie", "wol", "xho", "yor", "zho", "zul" ]

[ "### Human Annotations (Default)\n\n\nThe data fields are the same among all splits:\n\n\n* 'inputs': Prompt or input to the language model.\n* 'targets': Completion or output of the language model.\n* 'language': The language of the 'inputs' and 'targets'.\n* 'language\\_code': The ISO code for the language of the 'inputs' and 'targets'.\n* 'annotation\\_type': The value denoting whether 'inputs' and 'targets' are 'original\\_annotations' or 're-annotations'.\n* 'user\\_id': Unique identifier of the annotator who submitted the prompt-completion pair.", "### Demographics Data\n\n\nThe data fields are the same among all splits:\n\n\n* 'user\\_id': Unique identifier of the annotator who submitted the prompt-completion pair.\n* 'age\\_range': Age of the annotator. Ranges from 0 to 121.\n* 'gender': Gender of the annotator. The values are 'male', 'female', 'prefer not to say', 'non-binary' and 'others'.\n* 'languages': List of languages spoken by the annotator.\n* 'dialects': Dialects reported by the annotator.\n\n\nSome empty values may be represented as 'null'.\n\n\nData Splits\n-----------", "### Human Annotations (Default)\n\n\nThe following are the splits of the data:\n\n\nSplit: train, No. of instances: 202,364, Language Coverage: All\nSplit: test, No. of instances: 1,750, Language Coverage: 7 ('Standard Arabic', 'Yoruba', 'Turkish', 'English', 'Simplified Chinese', 'Portuguese', 'Telugu')", "### Demographics Data\n\n\nThe following are the splits of the data:\n\n\n\nData Instances\n--------------", "### Human Annotations (Default)\n\n\nAn example of 'train' looks as follows:", "### Demographics Data\n\n\nAn example of 'train' looks as follows:\n\n\nStatistics\n----------", "### Annotation Types\n\n\nThe following is the breakdown of original annotations and re-annotations in the final dataset.", "### Languages\n\n\nThe dataset covers 65 languages: 28 high-resource, 12 mid-resource, and 31 low-resource languages. The following is details about the languages, dialects & scripts included in the dataset.\n\n\n\n Languages Info \nISO Code: 'amh', Language: Amharic, Resources: Low\nISO Code: 'arb', 'ary', 'ars', 'acq', 'arz' & 'apc', Language: Arabic (Standard, Moroccan, Najdi, Ta'izzi-Adeni, Egyptian & South Levantine), Resources: High\nISO Code: 'ben', Language: Bengali, Resources: Mid\nISO Code: 'ceb', Language: Cebuano, Resources: Mid\nISO Code: 'dan', Language: Danish, Resources: Mid\nISO Code: 'deu', Language: German, Resources: High\nISO Code: 'ell', Language: Greek, Resources: Mid\nISO Code: 'eng', Language: English, Resources: High\nISO Code: 'eus', Language: Basque, Resources: High\nISO Code: 'fil', Language: Filipino, Resources: Mid\nISO Code: 'fin', Language: Finnish, Resources: Mid\nISO Code: 'fra', Language: French, Resources: High\nISO Code: 'gle', Language: Irish, Resources: Low\nISO Code: 'guj', Language: Gujarati, Resources: Low\nISO Code: 'hat', Language: Haitian Creole, Resources: Low\nISO Code: 'hau', Language: Hausa, Resources: Low\nISO Code: 'hin', Language: Hindi, Resources: High\nISO Code: 'hun', Language: Hungarian, Resources: High\nISO Code: 'ibo', Language: Igbo, Resources: Low\nISO Code: 'ind', Language: Indonesian, Resources: Mid\nISO Code: 'ita', Language: Italian, Resources: High\nISO Code: 'jav', Language: Javanese, Resources: Low\nISO Code: 'jpn', Language: Japanese, Resources: High\nISO Code: 'kan', Language: Kannada, Resources: Low\nISO Code: 'kir', Language: Kyrgyz, Resources: Low\nISO Code: 'kor', Language: Korean, Resources: Mid\nISO Code: 'kur', Language: Kurdish, Resources: Low\nISO Code: 'lit', Language: Lithuanian, Resources: Mid\nISO Code: 'mal', Language: Malayalam, Resources: Low\nISO Code: 'mar', Language: Marathi, Resources: Low\nISO Code: 'mlg', Language: Malagasy, Resources: Low\nISO Code: 'msa', Language: Malay, Resources: Mid\nISO Code: 'mya', Language: Burmese, Resources: Low\nISO Code: 'nep', Language: Nepali, Resources: Low\nISO Code: 'nld', Language: Dutch, Resources: High\nISO Code: 'nso', Language: Northern Sotho, Resources: Low\nISO Code: 'nya', Language: Chichewa, Resources: Low\nISO Code: 'pan', Language: Punjabi, Resources: Low\nISO Code: 'pes', Language: Persian, Resources: High\nISO Code: 'pol', Language: Polish, Resources: High\nISO Code: 'por', Language: Portuguese, Resources: High\nISO Code: 'pus', Language: Pashto, Resources: Low\nISO Code: 'rus', Language: Russian, Resources: High\nISO Code: 'sin', Language: Sinhala, Resources: Low\nISO Code: 'sna', Language: Shona, Resources: Low\nISO Code: 'snd', Language: Sindhi, Resources: Low\nISO Code: 'som', Language: Somali, Resources: Low\nISO Code: 'spa', Language: Spanish, Resources: High\nISO Code: 'sqi', Language: Albanian, Resources: Low\nISO Code: 'srp', Language: Serbian, Resources: High\nISO Code: 'sun', Language: Sundanese, Resources: Low\nISO Code: 'swa', Language: Swahili, Resources: Low\nISO Code: 'swe', Language: Swedish, Resources: High\nISO Code: 'tam', Language: Tamil, Resources: Mid\nISO Code: 'tel', Language: Telugu, Resources: Low\nISO Code: 'tha', Language: Thai, Resources: Mid\nISO Code: 'tur', Language: Turkish, Resources: High\nISO Code: 'ukr', Language: Ukrainian, Resources: Mid\nISO Code: 'urd', Language: Urdu, Resources: Mid\nISO Code: 'vie', Language: Vietnamese, Resources: High\nISO Code: 'wol', Language: Wolof, Resources: Low\nISO Code: 'xho', Language: Xhosa, Resources: Low\nISO Code: 'yor', Language: Yorùbá, Resources: Low\nISO Code: 'zho', Language: Chinese (Traditional & Simplified), Resources: High\nISO Code: 'zul', Language: Zulu, Resources: Low\n\n\n\n \n\nMotivations & Intentions\n========================\n\n\n* Curation Rationale: The curation effort employed an open-science approach to create a diverse instruction-style dataset through annotators across the globe that ensures comprehensive representation across all languages. The success of the curation effort, led by volunteers across diverse backgrounds, was significantly influenced by their hope to meaningfully bring NLP advancements to their languages.\n\n\nKnown Limitations\n=================\n\n\n* Language and dialect coverage: The dataset covers a limited fraction of the world's linguistic diversity, with 93% of languages not represented, facing challenges in distinguishing between languages and dialects, lacking coverage for many regional dialects, and excluding programming languages.\n* Uneven distribution of contributions: The dataset contains contributions in annotation activities, with a 'long tail' of annotators making only one or two contributions, leading to potential dataset imbalances across languages and a lack of diversity within certain language annotations.\n* Cultural and Personal Bias: In the dataset, certain languages have limited representation due to a few dominant annotators, potentially leading to a narrow viewpoint and skewed distribution of content, particularly towards certain domains like news.\n* Gendered Pronouns: Many of the languages in the Aya Dataset only contain pronouns that are explicitly gendered (e.g., Arabic) or that lack gender-neutral third-person pronouns for gender-neutral reference (e.g. Estonian).\n* Formality Distinctions: The dataset encompasses languages with diverse formality distinctions, involving honorifics and situational choices in pronoun use, reflecting varying levels of standardization influenced by regional, cultural, and identity factors.\n* Toxic or Offensive Speech: The Aya Annotation Platform lacked specific flags for toxic speech, relying on human verification and peer review to mitigate offensive content, but there's no guarantee that all potentially offensive data points were removed during the annotation process.\n* Accounting for mislabeled data: The Aya Annotation Platform lacks re-labeling capabilities, leading to potential mislabeled data in the Aya Dataset, including instances of incorrect language assignments and non-compliance with instruction-style formatting.\n\n\nAdditional Information\n======================\n\n\nProvenance\n----------\n\n\n* Methods Used: Crowd-sourced through volunteer annotations, followed by a quality assessment phase in which samples from the dataset were checked.\n* Methodology Details:\n\t+ *Source:* Original annotations and edits of opensource NLP datasets\n\t+ *Platform:* Aya Annotation Platform\n\t+ *Dates of Collection:* May 2023 - Dec 2023\n\n\nDataset Version and Maintenance\n-------------------------------\n\n\n* Maintenance Status: Actively Maintained\n* Version Details:\n\t+ *Current version:* 1.0\n\t+ *Last Update:* 02/2024\n\t+ *First Release:* 02/2024\n* Maintenance Plan: Updates will be periodically made available based on volunteer contributions.\n\n\nAuthorship\n----------\n\n\n* Publishing Organization: Cohere For AI\n* Industry Type: Not-for-profit - Tech\n* Contact Details: URL\n\n\nLicensing Information\n---------------------\n\n\nThis dataset can be used for any purpose, whether academic or commercial, under the terms of the Apache 2.0 License." ]

[ "TAGS\n#task_categories-other #annotations_creators-crowdsourced #annotations_creators-expert-generated #language_creators-crowdsourced #language_creators-expert-generated #multilinguality-multilingual #size_categories-100K<n<1M #source_datasets-original #language-Amharic #language-Standard Arabic #language-Moroccan Arabic #language-Najdi Arabic #language-Ta'izzi-Adeni Arabic #language-Egyptian Arabic #language-Levantine Arabic #language-Bengali #language-Cebuano #language-Danish #language-German #language-Modern Greek (1453-) #language-English #language-Basque #language-Filipino #language-Finnish #language-French #language-Irish #language-Gujarati #language-Haitian #language-Hausa #language-Hindi #language-Hungarian #language-Igbo #language-Indonesian #language-Italian #language-Javanese #language-Japanese #language-Kannada #language-Kirghiz #language-Korean #language-Kurdish #language-Lithuanian #language-Malayalam #language-Marathi #language-Malagasy #language-Malay (macrolanguage) #language-Burmese #language-Nepali (macrolanguage) #language-Dutch #language-Pedi #language-Nyanja #language-Panjabi #language-Iranian Persian #language-Polish #language-Portuguese #language-Pushto #language-Russian #language-Sinhala #language-Shona #language-Sindhi #language-Somali #language-Spanish #language-Albanian #language-Serbian #language-Sundanese #language-Swahili (macrolanguage) #language-Swedish #language-Tamil #language-Telugu #language-Thai #language-Turkish #language-Ukrainian #language-Urdu #language-Vietnamese #language-Wolof #language-Xhosa #language-Yoruba #language-Chinese #language-Zulu #license-apache-2.0 #arxiv-2402.06619 #region-us \n", "### Human Annotations (Default)\n\n\nThe data fields are the same among all splits:\n\n\n* 'inputs': Prompt or input to the language model.\n* 'targets': Completion or output of the language model.\n* 'language': The language of the 'inputs' and 'targets'.\n* 'language\\_code': The ISO code for the language of the 'inputs' and 'targets'.\n* 'annotation\\_type': The value denoting whether 'inputs' and 'targets' are 'original\\_annotations' or 're-annotations'.\n* 'user\\_id': Unique identifier of the annotator who submitted the prompt-completion pair.", "### Demographics Data\n\n\nThe data fields are the same among all splits:\n\n\n* 'user\\_id': Unique identifier of the annotator who submitted the prompt-completion pair.\n* 'age\\_range': Age of the annotator. Ranges from 0 to 121.\n* 'gender': Gender of the annotator. The values are 'male', 'female', 'prefer not to say', 'non-binary' and 'others'.\n* 'languages': List of languages spoken by the annotator.\n* 'dialects': Dialects reported by the annotator.\n\n\nSome empty values may be represented as 'null'.\n\n\nData Splits\n-----------", "### Human Annotations (Default)\n\n\nThe following are the splits of the data:\n\n\nSplit: train, No. of instances: 202,364, Language Coverage: All\nSplit: test, No. of instances: 1,750, Language Coverage: 7 ('Standard Arabic', 'Yoruba', 'Turkish', 'English', 'Simplified Chinese', 'Portuguese', 'Telugu')", "### Demographics Data\n\n\nThe following are the splits of the data:\n\n\n\nData Instances\n--------------", "### Human Annotations (Default)\n\n\nAn example of 'train' looks as follows:", "### Demographics Data\n\n\nAn example of 'train' looks as follows:\n\n\nStatistics\n----------", "### Annotation Types\n\n\nThe following is the breakdown of original annotations and re-annotations in the final dataset.", "### Languages\n\n\nThe dataset covers 65 languages: 28 high-resource, 12 mid-resource, and 31 low-resource languages. The following is details about the languages, dialects & scripts included in the dataset.\n\n\n\n Languages Info \nISO Code: 'amh', Language: Amharic, Resources: Low\nISO Code: 'arb', 'ary', 'ars', 'acq', 'arz' & 'apc', Language: Arabic (Standard, Moroccan, Najdi, Ta'izzi-Adeni, Egyptian & South Levantine), Resources: High\nISO Code: 'ben', Language: Bengali, Resources: Mid\nISO Code: 'ceb', Language: Cebuano, Resources: Mid\nISO Code: 'dan', Language: Danish, Resources: Mid\nISO Code: 'deu', Language: German, Resources: High\nISO Code: 'ell', Language: Greek, Resources: Mid\nISO Code: 'eng', Language: English, Resources: High\nISO Code: 'eus', Language: Basque, Resources: High\nISO Code: 'fil', Language: Filipino, Resources: Mid\nISO Code: 'fin', Language: Finnish, Resources: Mid\nISO Code: 'fra', Language: French, Resources: High\nISO Code: 'gle', Language: Irish, Resources: Low\nISO Code: 'guj', Language: Gujarati, Resources: Low\nISO Code: 'hat', Language: Haitian Creole, Resources: Low\nISO Code: 'hau', Language: Hausa, Resources: Low\nISO Code: 'hin', Language: Hindi, Resources: High\nISO Code: 'hun', Language: Hungarian, Resources: High\nISO Code: 'ibo', Language: Igbo, Resources: Low\nISO Code: 'ind', Language: Indonesian, Resources: Mid\nISO Code: 'ita', Language: Italian, Resources: High\nISO Code: 'jav', Language: Javanese, Resources: Low\nISO Code: 'jpn', Language: Japanese, Resources: High\nISO Code: 'kan', Language: Kannada, Resources: Low\nISO Code: 'kir', Language: Kyrgyz, Resources: Low\nISO Code: 'kor', Language: Korean, Resources: Mid\nISO Code: 'kur', Language: Kurdish, Resources: Low\nISO Code: 'lit', Language: Lithuanian, Resources: Mid\nISO Code: 'mal', Language: Malayalam, Resources: Low\nISO Code: 'mar', Language: Marathi, Resources: Low\nISO Code: 'mlg', Language: Malagasy, Resources: Low\nISO Code: 'msa', Language: Malay, Resources: Mid\nISO Code: 'mya', Language: Burmese, Resources: Low\nISO Code: 'nep', Language: Nepali, Resources: Low\nISO Code: 'nld', Language: Dutch, Resources: High\nISO Code: 'nso', Language: Northern Sotho, Resources: Low\nISO Code: 'nya', Language: Chichewa, Resources: Low\nISO Code: 'pan', Language: Punjabi, Resources: Low\nISO Code: 'pes', Language: Persian, Resources: High\nISO Code: 'pol', Language: Polish, Resources: High\nISO Code: 'por', Language: Portuguese, Resources: High\nISO Code: 'pus', Language: Pashto, Resources: Low\nISO Code: 'rus', Language: Russian, Resources: High\nISO Code: 'sin', Language: Sinhala, Resources: Low\nISO Code: 'sna', Language: Shona, Resources: Low\nISO Code: 'snd', Language: Sindhi, Resources: Low\nISO Code: 'som', Language: Somali, Resources: Low\nISO Code: 'spa', Language: Spanish, Resources: High\nISO Code: 'sqi', Language: Albanian, Resources: Low\nISO Code: 'srp', Language: Serbian, Resources: High\nISO Code: 'sun', Language: Sundanese, Resources: Low\nISO Code: 'swa', Language: Swahili, Resources: Low\nISO Code: 'swe', Language: Swedish, Resources: High\nISO Code: 'tam', Language: Tamil, Resources: Mid\nISO Code: 'tel', Language: Telugu, Resources: Low\nISO Code: 'tha', Language: Thai, Resources: Mid\nISO Code: 'tur', Language: Turkish, Resources: High\nISO Code: 'ukr', Language: Ukrainian, Resources: Mid\nISO Code: 'urd', Language: Urdu, Resources: Mid\nISO Code: 'vie', Language: Vietnamese, Resources: High\nISO Code: 'wol', Language: Wolof, Resources: Low\nISO Code: 'xho', Language: Xhosa, Resources: Low\nISO Code: 'yor', Language: Yorùbá, Resources: Low\nISO Code: 'zho', Language: Chinese (Traditional & Simplified), Resources: High\nISO Code: 'zul', Language: Zulu, Resources: Low\n\n\n\n \n\nMotivations & Intentions\n========================\n\n\n* Curation Rationale: The curation effort employed an open-science approach to create a diverse instruction-style dataset through annotators across the globe that ensures comprehensive representation across all languages. The success of the curation effort, led by volunteers across diverse backgrounds, was significantly influenced by their hope to meaningfully bring NLP advancements to their languages.\n\n\nKnown Limitations\n=================\n\n\n* Language and dialect coverage: The dataset covers a limited fraction of the world's linguistic diversity, with 93% of languages not represented, facing challenges in distinguishing between languages and dialects, lacking coverage for many regional dialects, and excluding programming languages.\n* Uneven distribution of contributions: The dataset contains contributions in annotation activities, with a 'long tail' of annotators making only one or two contributions, leading to potential dataset imbalances across languages and a lack of diversity within certain language annotations.\n* Cultural and Personal Bias: In the dataset, certain languages have limited representation due to a few dominant annotators, potentially leading to a narrow viewpoint and skewed distribution of content, particularly towards certain domains like news.\n* Gendered Pronouns: Many of the languages in the Aya Dataset only contain pronouns that are explicitly gendered (e.g., Arabic) or that lack gender-neutral third-person pronouns for gender-neutral reference (e.g. Estonian).\n* Formality Distinctions: The dataset encompasses languages with diverse formality distinctions, involving honorifics and situational choices in pronoun use, reflecting varying levels of standardization influenced by regional, cultural, and identity factors.\n* Toxic or Offensive Speech: The Aya Annotation Platform lacked specific flags for toxic speech, relying on human verification and peer review to mitigate offensive content, but there's no guarantee that all potentially offensive data points were removed during the annotation process.\n* Accounting for mislabeled data: The Aya Annotation Platform lacks re-labeling capabilities, leading to potential mislabeled data in the Aya Dataset, including instances of incorrect language assignments and non-compliance with instruction-style formatting.\n\n\nAdditional Information\n======================\n\n\nProvenance\n----------\n\n\n* Methods Used: Crowd-sourced through volunteer annotations, followed by a quality assessment phase in which samples from the dataset were checked.\n* Methodology Details:\n\t+ *Source:* Original annotations and edits of opensource NLP datasets\n\t+ *Platform:* Aya Annotation Platform\n\t+ *Dates of Collection:* May 2023 - Dec 2023\n\n\nDataset Version and Maintenance\n-------------------------------\n\n\n* Maintenance Status: Actively Maintained\n* Version Details:\n\t+ *Current version:* 1.0\n\t+ *Last Update:* 02/2024\n\t+ *First Release:* 02/2024\n* Maintenance Plan: Updates will be periodically made available based on volunteer contributions.\n\n\nAuthorship\n----------\n\n\n* Publishing Organization: Cohere For AI\n* Industry Type: Not-for-profit - Tech\n* Contact Details: URL\n\n\nLicensing Information\n---------------------\n\n\nThis dataset can be used for any purpose, whether academic or commercial, under the terms of the Apache 2.0 License." ]

[ "task_categories:text-classification", "task_categories:summarization", "task_categories:translation", "size_categories:100M<n<1B", "language:ace", "language:afr", "language:amh", "language:ara", "language:aze", "language:ban", "language:bbc", "language:bel", "language:bem", "language:ben", "language:bjn", "language:bul", "language:cat", "language:ceb", "language:ces", "language:cym", "language:dan", "language:deu", "language:ell", "language:eng", "language:epo", "language:est", "language:eus", "language:fil", "language:fin", "language:fon", "language:fra", "language:gla", "language:gle", "language:glg", "language:guj", "language:hat", "language:hau", "language:heb", "language:hin", "language:hrv", "language:hun", "language:hye", "language:ibo", "language:ind", "language:isl", "language:ita", "language:jav", "language:jpn", "language:kan", "language:kas", "language:kat", "language:kau", "language:kaz", "language:khm", "language:kin", "language:kir", "language:kor", "language:kur", "language:lao", "language:lav", "language:lij", "language:lit", "language:ltz", "language:mad", "language:mal", "language:man", "language:mar", "language:min", "language:mkd", "language:mlg", "language:mlt", "language:mon", "language:mri", "language:msa", "language:mya", "language:nep", "language:nij", "language:nld", "language:nor", "language:nso", "language:nya", "language:pan", "language:pes", "language:pol", "language:por", "language:pus", "language:ron", "language:rus", "language:sin", "language:slk", "language:slv", "language:smo", "language:sna", "language:snd", "language:som", "language:sot", "language:spa", "language:sqi", "language:srp", "language:sun", "language:swa", "language:swe", "language:tam", "language:taq", "language:tel", "language:tgk", "language:tha", "language:tur", "language:twi", "language:ukr", "language:urd", "language:uzb", "language:vie", "language:wol", "language:xho", "language:yid", "language:yor", "language:zho", "language:zul", "license:apache-2.0", "arxiv:2402.06619", "region:us" ]

[ "2402.06619" ]

[ "ace", "afr", "amh", "ara", "aze", "ban", "bbc", "bel", "bem", "ben", "bjn", "bul", "cat", "ceb", "ces", "cym", "dan", "deu", "ell", "eng", "epo", "est", "eus", "fil", "fin", "fon", "fra", "gla", "gle", "glg", "guj", "hat", "hau", "heb", "hin", "hrv", "hun", "hye", "ibo", "ind", "isl", "ita", "jav", "jpn", "kan", "kas", "kat", "kau", "kaz", "khm", "kin", "kir", "kor", "kur", "lao", "lav", "lij", "lit", "ltz", "mad", "mal", "man", "mar", "min", "mkd", "mlg", "mlt", "mon", "mri", "msa", "mya", "nep", "nij", "nld", "nor", "nso", "nya", "pan", "pes", "pol", "por", "pus", "ron", "rus", "sin", "slk", "slv", "smo", "sna", "snd", "som", "sot", "spa", "sqi", "srp", "sun", "swa", "swe", "tam", "taq", "tel", "tgk", "tha", "tur", "twi", "ukr", "urd", "uzb", "vie", "wol", "xho", "yid", "yor", "zho", "zul" ]

[ "### Statistics\n\n\nThe total number of data points, including the Aya Dataset' is 513,758,189. To view the breakdown of dialect codes and the respective templated and translated data point counts in the Aya Collection , refer to the toggled table below.\n\n\n\n **Breakdown of Aya Collection data point counts grouped by dialects** \n\nPS: Templated data also includes Mozambican Portuguese, which doesn't have its own ISO language code.\n\n\n\n \n\nMotivations & Intentions\n========================\n\n\n* Curation Rationale: Automatic augmentation of existing datasets serves to enhance the available linguistic resources for multiple languages. The list of languages was initially established from mT5 and aligned with the annotators’ language list and NLLB translation model. The datasets were translated directly from English for all languages.\n\n\nAdditional Information\n======================\n\n\nProvenance\n----------\n\n\n* Methods Used: A combination of crowd-sourced templating and automatic translation was employed to source this dataset.\n* Methodology Details:\n\t+ *Source:* Existing NLP datasets\n\t+ *Dates of Collection:* May 2023 - Dec 2023\n\n\nDataset Version and Maintenance\n-------------------------------\n\n\n* Maintenance Status: Actively Maintained\n* Version Details:\n\t+ *Current version:* 1.0\n\t+ *Last Update:* 02/2024\n\t+ *First Release:* 02/2024\n\n\nAuthorship\n----------\n\n\n* Publishing Organization: Cohere For AI\n* Industry Type: Not-for-profit - Tech\n* Contact Details: URL\n\n\nLicensing Information\n---------------------\n\n\nThis dataset can be used for any purpose, whether academic or commercial, under the terms of the Apache 2.0 License." ]

[ "TAGS\n#task_categories-text-classification #task_categories-summarization #task_categories-translation #size_categories-100M<n<1B #language-Achinese #language-Afrikaans #language-Amharic #language-Arabic #language-Azerbaijani #language-Balinese #language-Batak Toba #language-Belarusian #language-Bemba (Zambia) #language-Bengali #language-Banjar #language-Bulgarian #language-Catalan #language-Cebuano #language-Czech #language-Welsh #language-Danish #language-German #language-Modern Greek (1453-) #language-English #language-Esperanto #language-Estonian #language-Basque #language-Filipino #language-Finnish #language-Fon #language-French #language-Scottish Gaelic #language-Irish #language-Galician #language-Gujarati #language-Haitian #language-Hausa #language-Hebrew #language-Hindi #language-Croatian #language-Hungarian #language-Armenian #language-Igbo #language-Indonesian #language-Icelandic #language-Italian #language-Javanese #language-Japanese #language-Kannada #language-Kashmiri #language-Georgian #language-Kanuri #language-Kazakh #language-Khmer #language-Kinyarwanda #language-Kirghiz #language-Korean #language-Kurdish #language-Lao #language-Latvian #language-Ligurian #language-Lithuanian #language-Luxembourgish #language-Madurese #language-Malayalam #language-Mandingo #language-Marathi #language-Minangkabau #language-Macedonian #language-Malagasy #language-Maltese #language-Mongolian #language-Maori #language-Malay (macrolanguage) #language-Burmese #language-Nepali (macrolanguage) #language-Ngaju #language-Dutch #language-Norwegian #language-Pedi #language-Nyanja #language-Panjabi #language-Iranian Persian #language-Polish #language-Portuguese #language-Pushto #language-Romanian #language-Russian #language-Sinhala #language-Slovak #language-Slovenian #language-Samoan #language-Shona #language-Sindhi #language-Somali #language-Southern Sotho #language-Spanish #language-Albanian #language-Serbian #language-Sundanese #language-Swahili (macrolanguage) #language-Swedish #language-Tamil #language-Tamasheq #language-Telugu #language-Tajik #language-Thai #language-Turkish #language-Twi #language-Ukrainian #language-Urdu #language-Uzbek #language-Vietnamese #language-Wolof #language-Xhosa #language-Yiddish #language-Yoruba #language-Chinese #language-Zulu #license-apache-2.0 #arxiv-2402.06619 #region-us \n", "### Statistics\n\n\nThe total number of data points, including the Aya Dataset' is 513,758,189. To view the breakdown of dialect codes and the respective templated and translated data point counts in the Aya Collection , refer to the toggled table below.\n\n\n\n **Breakdown of Aya Collection data point counts grouped by dialects** \n\nPS: Templated data also includes Mozambican Portuguese, which doesn't have its own ISO language code.\n\n\n\n \n\nMotivations & Intentions\n========================\n\n\n* Curation Rationale: Automatic augmentation of existing datasets serves to enhance the available linguistic resources for multiple languages. The list of languages was initially established from mT5 and aligned with the annotators’ language list and NLLB translation model. The datasets were translated directly from English for all languages.\n\n\nAdditional Information\n======================\n\n\nProvenance\n----------\n\n\n* Methods Used: A combination of crowd-sourced templating and automatic translation was employed to source this dataset.\n* Methodology Details:\n\t+ *Source:* Existing NLP datasets\n\t+ *Dates of Collection:* May 2023 - Dec 2023\n\n\nDataset Version and Maintenance\n-------------------------------\n\n\n* Maintenance Status: Actively Maintained\n* Version Details:\n\t+ *Current version:* 1.0\n\t+ *Last Update:* 02/2024\n\t+ *First Release:* 02/2024\n\n\nAuthorship\n----------\n\n\n* Publishing Organization: Cohere For AI\n* Industry Type: Not-for-profit - Tech\n* Contact Details: URL\n\n\nLicensing Information\n---------------------\n\n\nThis dataset can be used for any purpose, whether academic or commercial, under the terms of the Apache 2.0 License." ]

[ "task_categories:text-classification", "task_categories:table-question-answering", "task_categories:question-answering", "task_categories:conversational", "size_categories:100K<n<1M", "license:mit", "code", "region:us" ]

[]

[]

[ "# AgoraX/AIEC-140K Dataset\n===============================\n\nExcited to Announce AgoraX/AIEC-140K!\n\nAn all-new dataset with super high High Quality AI Engineering Code Tokens totaling 140k samples!", "## Introduction\n------------\n\nThe AgoraX/AIEC-140K dataset is a collection of AI engineering code tokens from top research labs such as OpenAI, Nvidia, Google, Lucidrains, and others. These tokens have been scraped from various repositories on GitHub, providing a valuable resource for researchers and developers in the field of Artificial Intelligence.\n\nThis README file serves as a guide to understand the dataset and effectively utilize its contents.", "## Dataset Details\n---------------\n\n- Dataset Name: AgoraX/AIEC-140K\n- Total Samples: 140,000", "### Data Format\n\nThe dataset primarily consists of code tokens, which are the atomic units of code. Each code token is a single word or a character representing a meaningful entity in AI engineering code. These tokens were collected from different repositories, ensuring a diverse collection of samples.\n\nThe data does not include complete code snippets or files but focuses on individual tokens to enable easy integration and usage in various downstream tasks.", "### Data Sources\n\nCode tokens in the AgoraX/AIEC-140K dataset are scraped from various repositories on GitHub. Prominent research labs including OpenAI, Nvidia, Google, Lucidrains, and others have contributed to this dataset.\n\nPlease note that the dataset does not provide details on the exact repositories or sources from where each token is scraped.", "### Usage\n\nThe AgoraX/AIEC-140K dataset is a valuable resource for researchers, developers, and practitioners in the field of AI engineering. The dataset can be utilized for various purposes, including but not limited to:\n\n- Training language models for code generation\n- Pre-training and fine-tuning neural networks\n- Code completion and suggestion systems\n- Understanding and analyzing code patterns and trends in AI engineering\n\n--------\n\nIf you use the AgoraX/AIEC-140K dataset in your research work, please consider citing it using the following BibTeX:", "# License\n-------\n\nThe AgoraX/AIEC-140K dataset is released under the MIT License. Please refer to the LICENSE file in the dataset repository for more details.", "# Contact\n-------\n\nFor any further inquiries or feedback regarding the dataset, please contact the AgoraX Team in the discord: URL\n\nWe appreciate your interest and hope that the AgoraX/AIEC-140K dataset proves to be a valuable asset in advancing AI engineering research and development." ]

[ "TAGS\n#task_categories-text-classification #task_categories-table-question-answering #task_categories-question-answering #task_categories-conversational #size_categories-100K<n<1M #license-mit #code #region-us \n", "# AgoraX/AIEC-140K Dataset\n===============================\n\nExcited to Announce AgoraX/AIEC-140K!\n\nAn all-new dataset with super high High Quality AI Engineering Code Tokens totaling 140k samples!", "## Introduction\n------------\n\nThe AgoraX/AIEC-140K dataset is a collection of AI engineering code tokens from top research labs such as OpenAI, Nvidia, Google, Lucidrains, and others. These tokens have been scraped from various repositories on GitHub, providing a valuable resource for researchers and developers in the field of Artificial Intelligence.\n\nThis README file serves as a guide to understand the dataset and effectively utilize its contents.", "## Dataset Details\n---------------\n\n- Dataset Name: AgoraX/AIEC-140K\n- Total Samples: 140,000", "### Data Format\n\nThe dataset primarily consists of code tokens, which are the atomic units of code. Each code token is a single word or a character representing a meaningful entity in AI engineering code. These tokens were collected from different repositories, ensuring a diverse collection of samples.\n\nThe data does not include complete code snippets or files but focuses on individual tokens to enable easy integration and usage in various downstream tasks.", "### Data Sources\n\nCode tokens in the AgoraX/AIEC-140K dataset are scraped from various repositories on GitHub. Prominent research labs including OpenAI, Nvidia, Google, Lucidrains, and others have contributed to this dataset.\n\nPlease note that the dataset does not provide details on the exact repositories or sources from where each token is scraped.", "### Usage\n\nThe AgoraX/AIEC-140K dataset is a valuable resource for researchers, developers, and practitioners in the field of AI engineering. The dataset can be utilized for various purposes, including but not limited to:\n\n- Training language models for code generation\n- Pre-training and fine-tuning neural networks\n- Code completion and suggestion systems\n- Understanding and analyzing code patterns and trends in AI engineering\n\n--------\n\nIf you use the AgoraX/AIEC-140K dataset in your research work, please consider citing it using the following BibTeX:", "# License\n-------\n\nThe AgoraX/AIEC-140K dataset is released under the MIT License. Please refer to the LICENSE file in the dataset repository for more details.", "# Contact\n-------\n\nFor any further inquiries or feedback regarding the dataset, please contact the AgoraX Team in the discord: URL\n\nWe appreciate your interest and hope that the AgoraX/AIEC-140K dataset proves to be a valuable asset in advancing AI engineering research and development." ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"UC-first-turn-raj-tokenizer\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"UC-first-turn-raj-tokenizer\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for \"vp-er-14l\"\n\nMore Information needed" ]

[ "TAGS\n#region-us \n", "# Dataset Card for \"vp-er-14l\"\n\nMore Information needed" ]