# Log analysis is necessary for credible evaluation of AI agents

arXiv: https://arxiv.org/abs/2605.08545 · HTML: https://mlsanity.com/paper/2605.08545/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 32 references that are in the map.
- First posted on arXiv 8 May 2026.

## Builds on (most cited first, 9 of 9)

- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains](https://mlsanity.com/paper/2406.12045/) (2024)
- [ST-WebAgentBench: A Benchmark for Evaluating Safety and Trustworthiness in Web Agents](https://mlsanity.com/paper/2410.06703/) (2024)
- [Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluation](https://mlsanity.com/paper/2510.11977/) (2025)
- [Towards a Science of AI Agent Reliability](https://mlsanity.com/paper/2602.16666/) (2026)
- [SABER: Small Actions, Big Errors -- Safeguarding Mutating Steps in LLM Agents](https://mlsanity.com/paper/2512.07850/) (2025)
- [BrowserArena: Evaluating LLM Agents on Real-World Web Navigation Tasks](https://mlsanity.com/paper/2510.02418/) (2025)
- [Measuring AI Agents' Progress on Multi-Step Cyber Attack Scenarios](https://mlsanity.com/paper/2603.11214/) (2026)
- [Seven simple steps for log analysis in AI systems](https://mlsanity.com/paper/2604.09563/) (2026)

## Built on by (most cited first, 10 of 10)

- [Quantifying Overclaiming Propensity in Frontier LLM Agents](https://mlsanity.com/paper/2609.20812/) (2026)
- [Causal methods for LLM development and evaluation](https://mlsanity.com/paper/2605.25998/) (2026)
- [OpenCollab: A Multi-Agent Coding Framework with Programmable Collaboration and Controllable Runtime](https://mlsanity.com/paper/2609.38345/) (2026)
- [Evaluating Open-Weight E-Commerce Agents with Environment-Grounded Verification](https://mlsanity.com/paper/2609.16093/) (2026)
- [EnterpriseVal: Quantifying the Efficacy, Reliability and Value of Generative AI in the Enterprise](https://mlsanity.com/paper/2609.21841/) (2026)
- [RECLAIM: Can Agents Reproduce the Claims of Machine Learning Papers?](https://mlsanity.com/paper/2609.28850/) (2026)
- [What Can a Leaderboard Certify? Compositional Controllability for Fair Evaluation and Training of Biomedical Literature-Review Agents](https://mlsanity.com/paper/2609.32318/) (2026)
- [TraceDance: An Automated System for Building Agent Behavior Benchmarks from Real-World Agent Deployment Traces](https://mlsanity.com/paper/2609.33295/) (2026)
- [Open-Endedness Bench: Measuring Epistemic Process from Agent Records](https://mlsanity.com/paper/2610.02588/) (2026)
- [Bookkeeping, Composition, or Unreachable Gold? Reading MemoryAgentBench's Conflict-Resolution Scores Against a Frozen Last-Write Resolver](https://mlsanity.com/paper/2610.09193/) (2026)

## Abstract (arXiv)

Agent benchmarks typically report only final outcomes: pass or fail. This threatens evaluation credibility in three ways. First, scores may be inflated or deflated by shortcuts and benchmark artifacts, misrepresenting capability. Second, benchmark performance may fail to predict real-world utility due to scaffold limitations and recurring failure modes. Finally, capability scores may conceal dangerous or catastrophic actions taken by the agent. We argue that log analysis -- the systematic tracking and analysis of the inputs, execution, and outputs of an AI agent -- is necessary to overcome these validity threats and promote credible agent evaluation. In this paper, we (1) present a taxonomy of threats to credible evaluation documented through log analysis, and (2) develop a set of guiding principles for log analysis. We illustrate these principles on tau-Bench Airline, revealing that pass^5 performance was under-elicited by nearly 50% and surfacing deployment failure modes invisible to outcome metrics. We conclude with pragmatic recommendations to increase uptake of log analysis, directed at diverse stakeholders including benchmark creators, model developers, independent evaluators, and deployers.