Automated Transcript Analysis for Detecting Flaws in Agentic Benchmarks

• URL: http://arxiv.org/abs/2607.27518v1
• Date: Wed, 29 Jul 2026 23:17:08 GMT
• Title: Automated Transcript Analysis for Detecting Flaws in Agentic Benchmarks
• Abstract summary: We develop AI scanners to detect four types of validity issues: ground truth access, tool failure, guessing vulnerability, and answer format ambiguity.<n>Our scanners identify several verified quality issues in five widely used benchmarks, including cases unlikely to be caught by random manual inspection.
• Score: 2.4035041353107673
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Capabilities of frontier models are often assessed using agentic benchmarks. To trust these results, benchmarks must accurately measure what they claim to and be free from invalidating flaws. Previous manual audits of benchmarks such as SWE-Bench-Verified have uncovered several validity issues in transcripts. However, manual review is difficult to scale, and it is unclear whether automated methods can reliably surface flaws that compromise benchmark validity. In this paper, we developed AI scanners to detect four types of validity issues: ground truth access, tool failure, guessing vulnerability, and answer format ambiguity. We produced grading rubrics for each to instruct human labeling, and evaluated the scanners against human labels on a held-out test set of Inspect Evals benchmarks. Our scanners identified several verified quality issues in five widely used benchmarks, including cases unlikely to be caught by random manual inspection. Not all cases were identified, and scanner performance varied substantially across benchmarks, criteria and models. We highlight several open challenges to be addressed to improve scanners for stronger quality assurance claims, including broader standardization gaps in the evaluation field that degrade scanner performance. Together, these results serve as a proof of concept for using automated transcript analysis to audit benchmark quality more broadly.

Related papers

• Faults in Our Formal Benchmarking: Dataset Defects and Evaluation Failures in Lean Theorem Proving [13.87786274932945]
  Benchmarks for LLM-assisted theorem proving in Lean are often treated as intrinsically reliable because every solved instance comes with a machine-checked proof.<n>We audit five widely used Lean theorem-proving benchmarks and their forks, using corpus-scale static checkers to surface 4,833 findings.<n>We propose a fault taxonomy, a suite of automated checkers and recall-oriented semantic audit prompts, and release standards to guide the creation of formal math datasets.
  arXiv  Detail & Related papers  (2026-06-28T16:46:36Z)
• Automated Benchmark Auditing for AI Agents and Large Language Models [46.03841647776303]
  Auto Benchmark Audit (ABA) is an agentic framework that systematically audits individual benchmark tasks.<n>We run ABA on a collection of frontier LLM benchmarks and previous NeurIPS publications, totaling 168 benchmarks across nine domains.<n>ABA identifies critical issues including ambiguous task design, execution environment conflicts, and incorrect ground truths in over 25.7% of the evaluated tasks.
  arXiv  Detail & Related papers  (2026-05-25T17:44:21Z)
• BenchGuard: Who Guards the Benchmarks? Automated Auditing of LLM Agent Benchmarks [26.58983143152204]
  BenchGuard is the first automated auditing framework for task-oriented, execution-based agent benchmarks.<n>It identifies 12 author-confirmed issues in ScienceAgentBench and 83.3% of expert-identified issues on the BIXBench Verified-50 subset.<n>A full audit of 50 complex bioinformatics tasks costs under USD 15, making automated benchmark auditing a practical and valuable complement to human review.
  arXiv  Detail & Related papers  (2026-04-27T19:51:25Z)
• DeepFact: Co-Evolving Benchmarks and Agents for Deep Research Factuality [49.62610727661819]
  Existing fact-checkers are primarily designed for general-domain, factoid-style atomic claims.<n>We propose Audit-then-Score (AtS), where benchmark labels and rationales are explicitly revisable.<n>We instantiate AtS as DeepFact-Bench, a versioned DRR factuality benchmark with auditable rationales.
  arXiv  Detail & Related papers  (2026-03-06T05:05:57Z)
• BenchBrowser -- Collecting Evidence for Evaluating Benchmark Validity [29.31434780089624]
  High-level metadata is too coarse to convey the granular reality of benchmarks.<n>BenchBrowser surfaces evaluation items relevant to natural language use cases over 20 benchmark suites.<n>It helps quantify a critical gap between practitioner intent and what benchmarks actually test.
  arXiv  Detail & Related papers  (2026-02-25T04:02:36Z)
• A Coin Flip for Safety: LLM Judges Fail to Reliably Measure Adversarial Robustness [57.510025257780306]
  We show that existing validation protocols fail to account for substantial distribution shifts inherent to red-teaming.<n>We propose ReliableBench, a benchmark of behaviors that remain more consistently judgeable, and JudgeStressTest, a dataset designed to expose judge failures.
  arXiv  Detail & Related papers  (2026-02-04T15:13:35Z)
• PaperAudit-Bench: Benchmarking Error Detection in Research Papers for Critical Automated Peer Review [54.141490756509306]
  We introduce PaperAudit-Bench, which consists of two components: PaperAudit-Dataset, an error dataset, and PaperAudit-Review, an automated review framework.<n>Experiments on PaperAudit-Bench reveal large variability in error detectability across models and detection depths.<n>We show that the dataset supports training lightweight LLM detectors via SFT and RL, enabling effective error detection at reduced computational cost.
  arXiv  Detail & Related papers  (2026-01-07T04:26:12Z)
• Fantastic Bugs and Where to Find Them in AI Benchmarks [28.604919035475188]
  We introduce a framework for systematic benchmark revision that leverages statistical analysis of response patterns to flag potentially invalid questions.<n>Our approach builds on a core assumption commonly used in AI evaluations that the mean score sufficiently summarizes model performance.<n>Across nine widely used benchmarks, our method guides expert review to identify problematic questions with up to 84% precision.
  arXiv  Detail & Related papers  (2025-11-20T22:49:21Z)
• SVeritas: Benchmark for Robust Speaker Verification under Diverse Conditions [54.34001921326444]
  Speaker verification (SV) models are increasingly integrated into security, personalization, and access control systems.<n>Existing benchmarks evaluate only subsets of these conditions, missing others entirely.<n>We introduce SVeritas, a comprehensive Speaker Verification tasks benchmark suite, assessing SV systems under stressors like recording duration, spontaneity, content, noise, microphone distance, reverberation, channel mismatches, audio bandwidth, codecs, speaker age, and susceptibility to spoofing and adversarial attacks.
  arXiv  Detail & Related papers  (2025-09-21T14:11:16Z)
• Do Large Language Model Benchmarks Test Reliability? [66.1783478365998]
  We investigate how well current benchmarks quantify model reliability.<n>Motivated by this gap in the evaluation of reliability, we propose the concept of so-called platinum benchmarks.<n>We evaluate a wide range of models on these platinum benchmarks and find that, indeed, frontier LLMs still exhibit failures on simple tasks.
  arXiv  Detail & Related papers  (2025-02-05T18:58:19Z)
• DeepfakeBench: A Comprehensive Benchmark of Deepfake Detection [55.70982767084996]
  A critical yet frequently overlooked challenge in the field of deepfake detection is the lack of a standardized, unified, comprehensive benchmark. We present the first comprehensive benchmark for deepfake detection, called DeepfakeBench, which offers three key contributions. DeepfakeBench contains 15 state-of-the-art detection methods, 9CL datasets, a series of deepfake detection evaluation protocols and analysis tools, as well as comprehensive evaluations.
  arXiv  Detail & Related papers  (2023-07-04T01:34:41Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.