---
date: '2023-01-29T10:51:05'
hypothesis-meta:
  created: '2023-01-29T10:51:05.830376+00:00'
  document:
    title:
    - 2301.11305.pdf
  flagged: false
  group: __world__
  hidden: false
  id: 1KcQzJ_CEe26lwMlz8h5FA
  links:
    html: https://hypothes.is/a/1KcQzJ_CEe26lwMlz8h5FA
    incontext: https://hyp.is/1KcQzJ_CEe26lwMlz8h5FA/arxiv.org/pdf/2301.11305.pdf
    json: https://hypothes.is/api/annotations/1KcQzJ_CEe26lwMlz8h5FA
  permissions:
    admin:
    - acct:ravenscroftj@hypothes.is
    delete:
    - acct:ravenscroftj@hypothes.is
    read:
    - group:__world__
    update:
    - acct:ravenscroftj@hypothes.is
  tags:
  - chatgpt
  - detecting gpt
  target:
  - selector:
    - end: 39629
      start: 39349
      type: TextPositionSelector
    - exact: hile in this work, we use off-the-shelfmask-filling models such as T5
        and mT5 (for non-Englishlanguages), some domains may see reduced performanceif
        existing mask-filling models do not well represent thespace of meaningful
        rephrases, reducing the quality of thecurvature estimate.
      prefix: onable pertur-bation function. W
      suffix: ' While DetectGPT provides the be'
      type: TextQuoteSelector
    source: https://arxiv.org/pdf/2301.11305.pdf
  text: The approach requires access to language models that can meaningfully and
    accurately rephrase (perturbate) the outputs from the model under evaluation.
    If these things do not align then it may not work well.
  updated: '2023-01-29T10:51:05.830376+00:00'
  uri: https://arxiv.org/pdf/2301.11305.pdf
  user: acct:ravenscroftj@hypothes.is
  user_info:
    display_name: James Ravenscroft
in-reply-to: https://arxiv.org/pdf/2301.11305.pdf
tags:
- chatgpt
- detecting gpt
- hypothesis
type: annotation
url: /annotations/2023/01/29/1674989465

---



 <blockquote>hile in this work, we use off-the-shelfmask-filling models such as T5 and mT5 (for non-Englishlanguages), some domains may see reduced performanceif existing mask-filling models do not well represent thespace of meaningful rephrases, reducing the quality of thecurvature estimate.</blockquote>The approach requires access to language models that can meaningfully and accurately rephrase (perturbate) the outputs from the model under evaluation. If these things do not align then it may not work well.