Eric Mitchell commited on
Commit
b6c3ce7
·
1 Parent(s): a062447

Added caching notice.

Browse files
Files changed (1) hide show
  1. app.py +3 -1
app.py CHANGED
@@ -213,6 +213,8 @@ with gr.Blocks() as demo:
213
  gr.Markdown(
214
  """# Detecting GPT-2 Generations with DetectGPT
215
  ##### This web app is a demo of DetectGPT, described in [this paper](https://arxiv.org/abs/2301.11305). DetectGPT is a general-purpose method for using a language model to detect its own generations; **however, this proof-of-concept only detects if a particular piece of text came from [GPT-2](https://openai.com/blog/better-language-models/).** Detections on samples from other models may be particularly unreliable. We may add larger models like GPT-J (6B), GPT-NeoX (20B), or GPT-3 (175B) in the future; we perform evaluations with these and other models in our paper.
 
 
216
  ## Instructions
217
  ##### Enter some text in the text box at the bottom of the page and click the "Detect" button. You can try the example texts in the table below to get started, or use the generation box to generate your own text from GPT-2. We'd love to hear your thoughts (whether successes or failures) on DetectGPT at [[email protected]](mailto:[email protected])!
218
  #### This demo is experimental; its predictions should not be used to justify real-world decisions.
@@ -306,7 +308,7 @@ with gr.Blocks() as demo:
306
 
307
  This web app is a demo of the DetectGPT method described in [this paper](https://arxiv.org/pdf/2301.11305v1.pdf). We can't make any guarantees about the accuracy of the results, but we hope you find it interesting! We are very grateful for the [Ray](https://www.ray.io/) distributed compute framework for making this web app much, much easier to build.
308
 
309
- Privacy notice: this web app does not collect any personal information."""
310
  )
311
  with gr.Column(scale=1, min_width=70):
312
  pass
 
213
  gr.Markdown(
214
  """# Detecting GPT-2 Generations with DetectGPT
215
  ##### This web app is a demo of DetectGPT, described in [this paper](https://arxiv.org/abs/2301.11305). DetectGPT is a general-purpose method for using a language model to detect its own generations; **however, this proof-of-concept only detects if a particular piece of text came from [GPT-2](https://openai.com/blog/better-language-models/).** Detections on samples from other models may be particularly unreliable. We may add larger models like GPT-J (6B), GPT-NeoX (20B), or GPT-3 (175B) in the future; we perform evaluations with these and other models in our paper.
216
+ ##### This demo currently does not support languages using non-Latin script. Sorry for the inconvenience; we're hoping to add support soon!
217
+ ##### [Update 7 Mar 2023] Due to high traffic, we have begun caching requests locally. Please do not submit sensitive or private information to this demo.
218
  ## Instructions
219
  ##### Enter some text in the text box at the bottom of the page and click the "Detect" button. You can try the example texts in the table below to get started, or use the generation box to generate your own text from GPT-2. We'd love to hear your thoughts (whether successes or failures) on DetectGPT at [[email protected]](mailto:[email protected])!
220
  #### This demo is experimental; its predictions should not be used to justify real-world decisions.
 
308
 
309
  This web app is a demo of the DetectGPT method described in [this paper](https://arxiv.org/pdf/2301.11305v1.pdf). We can't make any guarantees about the accuracy of the results, but we hope you find it interesting! We are very grateful for the [Ray](https://www.ray.io/) distributed compute framework for making this web app much, much easier to build.
310
 
311
+ Privacy notice: this web app does not collect any personal information beyond the text you submit for detection, which is cached for performance reasons."""
312
  )
313
  with gr.Column(scale=1, min_width=70):
314
  pass