← Back to all articles
arXiv cs.AIOctober 7, 2026

Measurement-First Auditing of Agentic Leaderboards: Contamination Susceptibility, Matched-Control Re-evaluation, and Scorer Validation

Excerpt

arXiv:2610.05830v1 Announce Type: new Abstract: Agentic leaderboards increasingly evaluate systems on public benchmarks whose task statements and solution-bearing artifacts can remain accessible. We propose a measurement-first audit framework that distinguishes contamination claims according to the evidence required to support them. It separates three channels that require different evidence: training-time exposure, evaluation-time retrieval, and pipeline/scaffold leakage. Each channel is coded