Integrity report for Multiple Choice Questions: Reasoning Makes Large Language Models (LLMs) More Self-Confident, Especially When They are Wrong

A machine-verified record of the checks Pith has run against this paper: detector runs, findings, signed bundle events, and canonical identifiers.

arXiv:2501.09775

Paper page arXiv integrity.json

Detector runs

Findings

Signed record

The machine-readable record for this paper lives at /pith/2501.09775/integrity.json. Pith Number bundles also include signed pith.integrity.v1 events where a Pith Number exists.