metadata

language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:8760
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: nomic-ai/modernbert-embed-base
widget:
  - source_sentence: What is the interpretation described as inappropriate?
    sentences:
      - >-
        . Factors to be considered in determining the reasonableness of the
        lawyer’s expectation of confidentiality include the sensitivity of the
        information and the extent to which the privacy of the communication is
        protected by law or by a confidentiality agreement
      - >-
        . 20 of competition and rests on an inappropriate interpretation of SBA
        regulation 13 C.F.R. § 125.9(b)(3)(i). See SHS MJAR at 16–23; VCH MJAR
        at 16–23
      - >-
        . 29-2, the CIA’s declaration explains in much more detail what is meant
        by “intelligence sources and methods” or “intelligence activities,” see
        Third Lutz Decl. ¶–30
  - source_sentence: >-
      What is the source of the information regarding Senetas's knowledge about
      FDA approval?
    sentences:
      - >-
        . . . the exemption under which the deletion is made, shall be indicated
        at the place in the record where such deletion is made.” Id. Finally,
        the FOIA provides that “a court shall accord substantial weight to an
        affidavit of an agency concerning the agency’s determination as to
        technical feasibility under . . . subsection (b).” Id. § 552(a)(4)(B)
      - >-
        . 52 Senetas asserts that it learned about the plan to discontinue
        seeking FDA approval for DR’s products in September of 2018 after the
        decision had been made without any Board involvement. Galbally Dep. Tr.
        66:19-23
      - >-
        . Conclusion Video footage, like social media evidence, is susceptible
        to alteration, and the increased availability of new technology,
        particularly the advent of image-generating artificial intelligence, may
        present unique challenges in authenticating videos and photographs
  - source_sentence: What does Class Deviation CD-2020-14 allow for at the contract level?
    sentences:
      - >-
        social media company that 7At trial, the State had attempted to
        introduce evidence that was purportedly a printout from the MySpace page
        of the girlfriend of the defendant (whose nickname was allegedly
        “Boozy”) to demonstrate that the girlfriend had threatened a State’s
        witness
      - >-
        .” Supplement 2 to Class Deviation CD-2020-14 (Supplement 2), AR at
        2904. The Senior Procurement Executive further elaborated that Class
        Deviation CD-2020-14 “allowed for the use of ‘unpriced labor’ categories
        at the contract level for certain IDIQ multiple-award contracts.” Id
      - >-
        . Circuit has recognized that, separate from claims seeking relief for
        specific requests made under the FOIA, requesting parties may also
        assert a “claim that an agency policy or practice will impair the
        party’s lawful access to information in the future.” Payne Enters., Inc.
        v. United States, 837 F.2d 486, 491 (D.C. Cir. 1988) (emphasis in
        original); 31 accord Newport Aeronautical Sales v
  - source_sentence: >-
      What should the agency describe about the non-exempt material in a
      document?
    sentences:
      - >-
        . A straightforward reading of the 2019 NDAA reveals that the
        Commission’s members are “temporary” federal employees. The Commission
        “shall be considered . . . a temporary organization under [5 U.S.C. §
        3161].” Pub. L. No. 115-232, § 1051(a)(2). The Commission’s 15 members
        are “appointed for the life of the Commission” and are “Federal
        employees.” Id. § 1051(a)(4)(A), (6)–(7)
      - >-
        .15 Posteriormente, en armonía con el marco constitucional y doctrinario
        previamente reseñado, el 13 de julio de 2011, nuestra Legislatura
        aprobó, la Ley del Derecho sobre la Propia Imagen o Ley Núm. 139-201116.
        Dicho precepto legal estatuye una causa de acción en daños y perjuicios
        debido al uso no autorizado de la imagen con fines comerciales o
        publicitarios
      - >-
        . To this end, the Circuit has said that “[i]n addition to a statement
        of its reasons, an agency should also describe what proportion of the
        information in a document is non-exempt and how that material is
        dispersed throughout the document.” Id
  - source_sentence: Which offeror is mentioned as getting in if there is a points discrepancy?
    sentences:
      - >-
        . at 9:14–19 (“[I]f an offeror does not have the same number of points,
        if it’s the 130th offeror and it doesn’t have the same number of points
        as the 90th offeror, then the solicitation says the 90th offeror gets in
        and the 130th doesn’t.”)
      - >-
        . But the State had to establish that the communications were the
        handiwork of the defendant. It was in that context that temporal
        proximity came into play: The timing of the communications relative to
        other events connecting the defendant to the alleged crime was
        circumstantial evidence of the defendant’s authorship. Id. at 674-76
      - >-
        . Since the plaintiff does not address this issue in its sur-reply brief
        in No. 11-445, and because the plaintiff does not ask the Court to
        direct the DOJ to produce Document 3 to the plaintiff, the plaintiff
        does not appear to continue to challenge the DOJ’s decision to withhold
        Document 3. 140 recorded decision to implement the opinion.” Id. at 32
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy@1
  - cosine_accuracy@3
  - cosine_accuracy@5
  - cosine_accuracy@10
  - cosine_precision@1
  - cosine_precision@3
  - cosine_precision@5
  - cosine_precision@10
  - cosine_recall@1
  - cosine_recall@3
  - cosine_recall@5
  - cosine_recall@10
  - cosine_ndcg@10
  - cosine_mrr@10
  - cosine_map@100
model-index:
  - name: Fine-tuned with [QuicKB](https://github.com/ALucek/QuicKB)
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 768
          type: dim_768
        metrics:
          - type: cosine_accuracy@1
            value: 0.582135523613963
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7494866529774127
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.795687885010267
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8572895277207392
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.582135523613963
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.24982888432580422
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1591375770020534
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08572895277207392
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.582135523613963
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7494866529774127
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.795687885010267
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8572895277207392
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7211793259435271
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6775296600501939
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6827316333877884
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 512
          type: dim_512
        metrics:
          - type: cosine_accuracy@1
            value: 0.5657084188911704
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7330595482546202
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7915811088295688
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8531827515400411
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5657084188911704
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.24435318275154005
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.15831622176591376
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08531827515400411
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.5657084188911704
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7330595482546202
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7915811088295688
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8531827515400411
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7102670568981261
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6645362765229291
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6695389256684248
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 256
          type: dim_256
        metrics:
          - type: cosine_accuracy@1
            value: 0.5410677618069816
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7063655030800822
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7659137577002053
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8305954825462012
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5410677618069816
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2354551676933607
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.15318275154004105
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08305954825462013
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.5410677618069816
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7063655030800822
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7659137577002053
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8305954825462012
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6839216686374571
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6371842508392814
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6427516419970609
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 128
          type: dim_128
        metrics:
          - type: cosine_accuracy@1
            value: 0.4887063655030801
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.6581108829568788
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7176591375770021
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7802874743326489
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.4887063655030801
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2193702943189596
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.14353182751540042
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07802874743326488
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.4887063655030801
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.6581108829568788
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7176591375770021
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7802874743326489
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6318826024721981
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.5846004041589256
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.5917468903182894
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 64
          type: dim_64
        metrics:
          - type: cosine_accuracy@1
            value: 0.3798767967145791
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5462012320328542
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.6139630390143738
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.704312114989733
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.3798767967145791
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1820670773442847
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.12279260780287474
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.0704312114989733
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.3798767967145791
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.5462012320328542
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.6139630390143738
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.704312114989733
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.5333651837657117
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.4796983475114887
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.4877644055271696
            name: Cosine Map@100

Fine-tuned with QuicKB

This is a sentence-transformers model finetuned from nomic-ai/modernbert-embed-base. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: nomic-ai/modernbert-embed-base
Maximum Sequence Length: 1024 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Language: en
License: apache-2.0

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 1024, 'do_lower_case': False}) with Transformer model: ModernBertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("AdamLucek/modernbert-embed-quickb")
# Run inference
sentences = [
    'Which offeror is mentioned as getting in if there is a points discrepancy?',
    '. at 9:14–19 (“[I]f an offeror does not have the same number of points, if it’s the 130th offeror and it doesn’t have the same number of points as the 90th offeror, then the solicitation says the 90th offeror gets in and the 130th doesn’t.”)',
    '. Since the plaintiff does not address this issue in its sur-reply brief in No. 11-445, and because the plaintiff does not ask the Court to direct the DOJ to produce Document 3 to the plaintiff, the plaintiff does not appear to continue to challenge the DOJ’s decision to withhold Document 3. 140 recorded decision to implement the opinion.” Id. at 32',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Information Retrieval

Datasets: dim_768, dim_512, dim_256, dim_128 and dim_64
Evaluated with InformationRetrievalEvaluator

Metric	dim_768	dim_512	dim_256	dim_128	dim_64
cosine_accuracy@1	0.5821	0.5657	0.5411	0.4887	0.3799
cosine_accuracy@3	0.7495	0.7331	0.7064	0.6581	0.5462
cosine_accuracy@5	0.7957	0.7916	0.7659	0.7177	0.614
cosine_accuracy@10	0.8573	0.8532	0.8306	0.7803	0.7043
cosine_precision@1	0.5821	0.5657	0.5411	0.4887	0.3799
cosine_precision@3	0.2498	0.2444	0.2355	0.2194	0.1821
cosine_precision@5	0.1591	0.1583	0.1532	0.1435	0.1228
cosine_precision@10	0.0857	0.0853	0.0831	0.078	0.0704
cosine_recall@1	0.5821	0.5657	0.5411	0.4887	0.3799
cosine_recall@3	0.7495	0.7331	0.7064	0.6581	0.5462
cosine_recall@5	0.7957	0.7916	0.7659	0.7177	0.614
cosine_recall@10	0.8573	0.8532	0.8306	0.7803	0.7043
cosine_ndcg@10	0.7212	0.7103	0.6839	0.6319	0.5334
cosine_mrr@10	0.6775	0.6645	0.6372	0.5846	0.4797
cosine_map@100	0.6827	0.6695	0.6428	0.5917	0.4878

Training Details

Training Dataset

Unnamed Dataset

Size: 8,760 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 7 tokens
mean: 15.54 tokens
max: 32 tokens

min: 4 tokens
mean: 76.24 tokens
max: 169 tokens

	anchor	positive
type	string	string
details	min: 7 tokens mean: 15.54 tokens max: 32 tokens	min: 4 tokens mean: 76.24 tokens max: 169 tokens

Samples:

anchor	positive
`What is being compared in the Circuit's statement?`	`.2d at 1389–90. The Circuit rejected this analogy, stating that, in contrast to the CIA Act, the NSA Act “protects not only organizational matters . . . but also ‘any information with respect to the activities’ of the NSA.” Id. at 1390`
`What type of internal documents used by the CIA in FOIA requests is mentioned?`	`. 108 Accordingly, the Court holds that certain specific categories of information withheld by the CIA in this case pursuant to § 403g clearly fall outside that provision’s scope, including (1) internal templates utilized by the CIA in tasking FOIA requests, (2) internal rules, policies and procedures governing FOIA processing, and (7) information about the CIA’s “core functions,” including`
`How many documents did the CIA withhold under Exemption 2?`	`. The CIA states in its declaration that all thirteen documents withheld under 38 The plaintiff previously indicated that it intended to challenge Exemption 2 withholding decisions made by the ODNI as well. See Hackett Decl. Ex. E at 1, ECF No. 29-8. The plaintiff, however, does not pursue that challenge in its opposition to the defendants’ motions for summary judgment in No. 11-445`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        768,
        512,
        256,
        128,
        64
    ],
    "matryoshka_weights": [
        1,
        1,
        1,
        1,
        1
    ],
    "n_dims_per_step": -1
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 32
gradient_accumulation_steps: 16
learning_rate: 2e-05
num_train_epochs: 4
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: True
tf32: True
load_best_model_at_end: True
optim: adamw_torch_fused
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 32
per_device_eval_batch_size: 8
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 16
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 2e-05
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 4
max_steps: -1
lr_scheduler_type: cosine
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: True
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: True
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch_fused
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
dataloader_pin_memory: True
dataloader_persistent_workers: False
skip_memory_metrics: True
use_legacy_prediction_loop: False
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: None
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
ray_scope: last
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
dispatch_batches: None
split_batches: None
include_tokens_per_second: False
include_num_input_tokens_seen: False
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_768_cosine_ndcg@10	dim_512_cosine_ndcg@10	dim_256_cosine_ndcg@10	dim_128_cosine_ndcg@10	dim_64_cosine_ndcg@10
0.5839	10	67.1727	-	-	-	-	-
1.0	18	-	0.6999	0.6820	0.6577	0.5988	0.4855
1.1168	20	32.4667	-	-	-	-	-
1.7007	30	27.9435	-	-	-	-	-
2.0	36	-	0.7167	0.7002	0.6764	0.6233	0.5187
2.2336	40	22.2924	-	-	-	-	-
2.8175	50	20.5125	-	-	-	-	-
3.0	54	-	0.7190	0.7080	0.6824	0.6318	0.5339
3.3504	60	18.3621	-	-	-	-	-
3.8175	68	-	0.7212	0.7103	0.6839	0.6319	0.5334

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.4.0
Transformers: 4.48.1
PyTorch: 2.5.1+cu124
Accelerate: 1.3.0
Datasets: 3.2.0
Tokenizers: 0.21.0

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}