Review papers using highlights and go!
AnnotateGPT is a chrome extension to allow conference and journal reviewers to review papers using annotations assisted by LLM such as GPT-4 and Claude 2.
Main features:
AutoAnnotating. LLM-generated annotations as a starting point for reviewers, streamlining the review process by focusing on relevant excerpts based on specific criteria.
Color-coding highlighting. You define a Review Model (e.g., originality, legibility and so on), each model’s attribute is mapped to a colour to be used during highlighting at review time,
Qualify highlighting. Highlights can be associated with comments, grades (strengths and weaknesses) or references to the literature. Your comments would undertake a sentiment analysis to avoid offensive wordings.
Canvas view. Have a global picture of the review so far. The canvas is plotted along with the attributes of the review model. Gradations and highlights are shown within each plot.
Review-draft generation. A first text draft is generated as a review head-start. Comments are placed by the manuscript quotes for authors to easily spot the rationales for the reviewer comments.
Sharing. Data is stored locally. Yet, it can be exported as a JSON file and emailed to colleagues who can then import it into their AnnotateGPT installations. On loading the manuscript, your colleagues will see the very same view as you.