BenchGuard: Who Guards the Benchmarks? Automated Auditing of LLM Agent Benchmarks

• URL: http://arxiv.org/abs/2604.24955v1
• Date: Mon, 27 Apr 2026 19:51:25 GMT
• Title: BenchGuard: Who Guards the Benchmarks? Automated Auditing of LLM Agent Benchmarks
• Abstract summary: 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.
• Score: 26.58983143152204
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As benchmarks grow in complexity, many apparent agent failures are not failures of the agent at all - they are failures of the benchmark itself: broken specifications, implicit assumptions, and rigid evaluation scripts that penalize valid alternative approaches. We propose employing frontier LLMs as systematic auditors of evaluation infrastructure, and realize this vision through BenchGuard, the first automated auditing framework for task-oriented, execution-based agent benchmarks. BenchGuard cross-verifies all benchmark artifacts via structured LLM protocols, optionally incorporating agent solutions or execution traces as additional diagnostic evidence. Deployed on two prominent scientific benchmarks, BenchGuard identified 12 author-confirmed issues in ScienceAgentBench - including fatal errors rendering tasks unsolvable - and exactly matched 83.3% of expert-identified issues on the BIXBench Verified-50 subset, catching defects that prior human review missed entirely. A full audit of 50 complex bioinformatics tasks costs under USD 15, making automated benchmark auditing a practical and valuable complement to human review. These findings point toward AI-assisted benchmark development, where frontier models serve not only as subjects of evaluation but as active participants in validating the evaluation infrastructure itself.

Related papers

• Benchmarking the Benchmarks: A Validity Audit of Tool-Calling Evaluation [3.079997053226693]
  We present a systematic validity and audit of four major tool-calling benchmark families.<n>Across 496 expert-reviewed benchmark tasks, we find 92 evaluator-human disagreements, corresponding to an 18.5% misalignment rate.<n>We introduce a unified taxonomy of tool-calling evaluation failures, release trace-level audit artifacts and corrected evaluation components, and argue for metrics that separately measure tool invocation, task completion, and outcome verification.
  arXiv  Detail & Related papers  (2026-06-30T21:10:42Z)
• TADDLE: A Tool-Augmented Agent for Detecting Deficient LLM-Generated Peer Reviews [4.404176366651557]
  We introduce TADDLE, a tool for detecting deficiencies in LLM-generated peer reviews.<n>Our benchmark comprises 1,800 reviews on 50 ICLR 2025 papers, multi-label-annotated by 18 domain experts.<n>Extensive experiments show that TADDLE performs strongly on both binary detection and the multi-label classification task.
  arXiv  Detail & Related papers  (2026-05-26T12:07:44Z)
• 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)
• Time to REFLECT: Can We Trust LLM Judges for Evidence-based Research Agents? [61.49434544687523]
  We introduce REFLECT, a meta-evaluation benchmark targeting fine-grained failure detection in agentic environments.<n> REFLECT defines a detailed taxonomy of process- and outcome-level failure modes, instantiated by performing controlled and localized interventions.<n>Our experiments show that even the best-performing models achieve overall accuracies below 55% across reasoning, tool-use, and report-quality failures.
  arXiv  Detail & Related papers  (2026-05-18T23:55:08Z)
• 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)
• Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluation [87.47155146067962]
  We provide a standardized evaluation harness that orchestrates parallel evaluations across hundreds of tasks.<n>We conduct three-dimensional analysis spanning models, scaffolds, and benchmarks.<n>Our analysis reveals surprising insights, such as higher reasoning effort reducing accuracy in the majority of runs.
  arXiv  Detail & Related papers  (2025-10-13T22:22:28Z)
• Establishing Best Practices for Building Rigorous Agentic Benchmarks [94.69724201080155]
  We show that many agentic benchmarks have issues in task setup or reward design.<n>Such issues can lead to under- or overestimation of agents' performance by up to 100% in relative terms.<n>We introduce the Agentic Benchmark Checklist (ABC), a set of guidelines that we synthesized from our benchmark-building experience.
  arXiv  Detail & Related papers  (2025-07-03T17:35:31Z)
• AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories [61.38499597241457]
  We propose AgentRewardBench, the first benchmark to assess the effectiveness of LLM judges for evaluating web agents.<n>Using our benchmark, we evaluate 12 LLM judges and find that no single LLM excels across all benchmarks.<n>We also find that the rule-based evaluation used by common benchmarks tends to underreport the success rate of web agents.
  arXiv  Detail & Related papers  (2025-04-11T19:49:22Z)

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.