paper-claim-audit

To verify numeric claims against raw data, this audit extracts quantitative statements from manuscripts and maps them to original CSV or JSON result files, producing a structured report confirming whether each number exactly matches the evidence.

A zero-context reviewer audits paper claims using a fresh model thread with no prior context, preventing confirmation bias during the verification of quantitative claims against raw data files.

To audit a manuscript directory, provide access to the raw result files in formats like CSV or JSON, and the audit will extract numeric claims from the paper to verify them against the evidence while generating a structured JSON and human-readable report.

Yes, this audit applies to both pre-submission checks and post-review audits, verifying that every numeric claim in a manuscript with associated data files and tables exactly matches the raw results.

Verifying research paper reproducibility requires access to raw result files such as CSV or JSON, stored in a dedicated paper directory alongside the manuscript, to map and validate quantitative claims against the evidence.

Yes, the audit enforces strict separation between claims and evidence by excluding executor summaries, ensuring that the zero-context reviewer validates numeric claims solely against raw data files without bias.