CORTEXA
← Browse
arxivcs.LGcs.SE2026-07-01

Auditing the Audit: Five Failure Modes in Benchmark-Validity Audits

Yanhang Li, Zhichao Fan, Zexin Zhuang

Governance frameworks ask AI providers and auditors for documented evaluation evidence, and perturbation-based construct-validity audits are a common form of that evidence. We argue the audits are themselves fragile: their conclusions can be silently manufactured by implementation details that readers cannot see in the reported numbers. We name five classes of pipeline failure and demonstrate each in a self-audit over safety benchmarks and open-weight instruction-tuned models. Under a unified six-point due-diligence gate, every cell lands in a non-confirmatory bucket, and no cell reaches confirmatory. The evidence here is a single two-model, five-benchmark case study, and F1--F5 is an illustrative, deliberately non-exhaustive starting taxonomy -- not a comprehensive partition of audit failures. We position the gate as a withholding and disclosure protocol for assurance-grade evidence, supplementary to (not a replacement for) classical construct-validity evidence, and not as a route to benchmark-validity verdicts.

View free PDFSource page

Related papers

arxivcs.AIcs.CLcs.LGcs.SE2026-06-29

EvalSafetyGap: A Hybrid Survey and Conceptual Framework for LLM Evaluation-Safety Failures

Buğra Alperen Uluırmak, Rifat Kurban

LLM evaluation and AI safety face a shared measurement problem: benchmark scores, reward-model signals, and reported safety metrics can improve while the latent properties they are meant to represent remain difficult to verify. This paper combines a hybrid survey - a systematic s…

View free PDFSource page
arxivcs.SEcs.AIcs.LGcs.PF2026-07-16

AEVAL: From Anecdotal to Deterministic Testing for Agentic Skill Workflows

Tejas Singh Anand, Yuet Ying Christina Wang, Wanting Jiang, Steve Masson, Tian Zheng, Bingjie Zhou

Modern agentic systems increasingly rely on skills: installable packages of natural language and code that teach an LLM agent to perform a domain task. As skill repositories grow, developers need automated quality signals on every change, yet evaluation today is largely anecdotal…

View free PDFSource page
arxivcs.SEcs.AIcs.LG2026-07-01

GPUAlert: A Zero-Instrumentation Process-Boundary Monitor for Diagnosing GPU Training-Job Failures

Parv Agarwal, Asif Ekbal

GPU training jobs fail often, roughly two in five on large production clusters, yet the operator typically learns of a failure only by reconnecting hours later. Experiment trackers require editing the training script and maintaining a cloud connection; the scheduler's mail hook d…

View free PDFSource page