Critique a research paper with AI
Check the central claims against the evidence, distinguish consequential weaknesses from minor faults and identify what would change the conclusion.
To critique a research paper with AI, start with a claim that matters to its contribution. A new method may score higher while benefiting from a larger tuning budget than its baseline. The paper-critique guide asks where that difference appears and how much it affects confidence in the comparison.
Different defects have different consequences. A typographical error may leave the reported result intact; a confounded evaluation can undermine the main claim. Request specific sections, figures or tables for each concern. Later evidence can add context, but its findings should remain distinct from the original paper’s results.
What it does
- Match a claim with its supporting evidence
- Check whether comparison conditions are fair
- Assess scope and possible threats to validity
- Identify missing details needed for reproduction
- Look for later work that tests a central claim
Install and use
- Install the skill and provide an authorised paper. Web access is needed to check later studies; an available full text is needed for location-specific criticism.
- Ask “Identify three central claims. For each, cite the evidence and assess whether the comparison is fair enough to support the stated conclusion.”
- Verify the cited passages and judge the work against the standards and tools available at publication.
FAQ
Source
Made by Jacquard Labs. Original skill: paper-critique, published on GitHub under the MIT licence.
View on GitHub