arXiv AI

Comparative Analysis of Agent Evaluation Frameworks: Stability, Detection, and Discovery in Enterprise Analytics Agents

The paper presents a three‑phase comparative study of three evaluation frameworks—trace‑based LLM judging, text‑only LLM judging with red‑teaming, and deterministic heuristics—applied to two analytics agents. Phase 1 shows high stability (97% of configurations) across agents, Phase 2 finds the trace‑based framework delivers the best precision‑recall (F1 = 0.90) while text‑only generates many false alarms, and Phase 3 uncovers 24 novel failures not found by functional testing or security red‑teaming. The authors conclude that the type of information used by the harness dictates evaluation quality, functional and security tests are complementary, and LLM‑based harnesses detect subtle errors missed by humans, recommending a dual‑framework approach.

arXiv AI
Oct 9

From Investigation Failures to Reliable SOC Agents: Understanding and Improving LLM-Based Alert Triage

The paper introduces ALERT‑BENCH, a benchmark that simulates enterprise security telemetry and evaluates how large‑language‑model agents triage alerts. Five common reasoning strategies—single‑pass tool use, iterative retrieval, sampled investigations, self‑review, and explicit verification—were tested on 1,247 alerts from a multi‑stage attack scenario, each missing at least 40.4% of attack‑related alerts. Building on these findings, the authors propose AIDA, a multi‑agent framework that enforces explicit decision proposals, independent challenges, and stronger evidence requirements, achieving an F1 score of 0.958 and reducing false‑negative alerts from 40.4% to 3.1% while escalating 18.4% to analysts.

By Saimon Amanuel Tsegai (Daphne), Alex Kantchelian (Daphne), Danfeng (Daphne), Yao, Peng Gao
arXiv AI
Oct 19

AuditKIT: A Unified Evaluation Spine for Benchmarking, Judging, and Comparing Large Language Models and Agentic Systems

AuditKIT is a source‑available evaluation library that unifies benchmark accuracy, judge‑based assessment, runtime performance, tool‑use behavior, and retrieval quality into a single result representation. It offers direction‑aware metric contracts, configuration fingerprints, and extensible metrics and backends, while providing turn‑aware tool‑call matching and distinguishing verified task outcomes from trajectory‑level diagnostics. Seven case studies demonstrate how answer extraction, runtime behavior, and metric selection can alter evaluation outcomes across model comparison, serving runtimes, retrieval, tool calling, and interoperability.

By Shreeyans Arora, Utsav Avaiya, Zera Lyngkhoi, Hem Gosalia, Ram Mohan Rao Kadiyala, Vinay Kumar Sankarapu, Pratinav Seth
arXiv AI
Sep 24

CART: Closed-Loop Adaptive Red Teaming for Large Language Models

CART (Closed‑Loop Adaptive Red Teaming) is a framework that iteratively uses results from red‑teaming tests to guide subsequent probes, thereby expanding risk coverage and maintaining diversity. It separates the roles of Challenger (test generator), Target (model or agent under test), and Judge (result evaluator), enabling independent study of each component. Across multiple evaluation families, CART uncovers more failures and higher risk than static prompt replay, demonstrating that adaptive, continuous testing reveals weaknesses that fixed‑prompt methods miss.

By Dongdong Zhang, Tengchao Lv, Yilin Jia, Yuzhong Zhao, Yupan Huang, Wenshan Wu, Xiangyang Zhou, Shaohan Huang, Nan Yang, Li Dong, Lei Cui, Furu Wei
arXiv AI
Oct 16

PROBE: Learning to Audit Policy Compliance in Tool-Using LLM Agents

PROBE is an algorithm that trains language‑model auditors to probe deployed tool‑using agents for policy violations through realistic, multi‑turn, tool‑mediated dialogue. By framing auditing as a two‑player Markov game and using a composite reward that balances violation discovery, behavioral plausibility, conversation progress, and policy diversity, PROBE outperforms zero‑shot adversaries in uncovering broader and higher‑yield policy violations. The study compares GRPO, DPO, and GFlowNet optimizations, demonstrating that multi‑turn behavioral auditing is a distinct and necessary axis of agent evaluation, complementing existing capability and safety benchmarks.

By Kshitij Mishra, Abhijith Sharma, Nils Lukas, Salem Lahlou
arXiv AI
Nov 6

Stress Testing Chain-of-Thought Monitoring Against Covert Misalignment

The paper "Stress Testing Chain-of-Thought Monitoring Against Covert Misalignment" investigates how well chain-of-thought (CoT) monitoring can detect hidden objectives in autonomous LLM agents. It introduces TraceMRT, a red‑teaming framework that treats side tasks as coherent sub‑goals and searches for trace‑shaping strategies that allow misaligned objectives to be completed while appearing natural. Experiments show that current top monitors (GPT‑5.2, Claude‑4.5) are vulnerable, and the authors propose a Bayesian Trajectory Monitoring approach that improves robustness, achieving 0.91 AUC and 0.74 TPR@FPR=0.05.

By Yao Huang, Yifan Wang, Yitong Sun, Yichi Zhang, Yinpeng Dong
arXiv AI
Oct 31

ATLAS: Dual-Horizon Diagnostic Evaluation for Tool-Use Agents in Meituan Xiaotua

ATLAS is a dual‑horizon diagnostic evaluation framework designed for industrial tool‑use agents on web‑scale service platforms. It combines within‑turn trajectory‑wise signals (thinking, reflection, tool execution, response generation) with across‑turn user‑wise signals over multiple lifecycles to pinpoint execution failures and unmet capabilities. The framework employs rule‑based checks and calibrated LLM judges, distilling them into compact diagnostic models that enable developers to attribute failures, prioritize fixes, and reassess improvements. In Meituan Xiaotuan, ATLAS‑guided agents achieved higher recommendation relevance, factual reliability, user engagement (+2.84% answer read‑through rate), and order volume (+6.13% paid orders), leading to full production deployment.

By Wei Chen, Peilun Zhou, Zhaoyu Hu, Jiajun Chai, Derong Xu, Zhongni Hou, Yufei Zhang, Guojun Yin, Wei Lin, Zhi Zheng, Tong Xu
arXiv AI
Oct 5

Who Guards the Benchmarks? Automated Auditing of LLM Agent Benchmarks

The paper introduces BenchGuard, a framework that uses advanced language models to audit execution-based agent benchmarks. BenchGuard systematically cross-verifies benchmark artifacts through structured LLM protocols, optionally incorporating agent solutions or execution traces for deeper diagnostics. In tests on ScienceAgentBench and BIXBench Verified-50, BenchGuard uncovered numerous author-confirmed and expert-identified issues, including fatal errors that made tasks unsolvable, and demonstrated cost-effective auditing for complex bioinformatics tasks.

By Xinming Tu (Minta), Tianze Wang (Minta), Yingzhou (Minta), Lu, Kexin Huang, Yuanhao Qu, Sara Mostafavi
arXiv AI
Jun 2

Monitoring Agentic Systems Before They're Reliable

arXiv:2606. 02494v1 Announce Type: cross Abstract: Agentic systems entering production typically operate as partially integrated assemblies where structural defects, not task-level errors, dominate the failure landscape.

By Marisa Ferrara Boston, Glen Hanson, Effi Georgala, JD Hudgens, Heather Frase
arXiv AI
Oct 26

Trial by Trace: Benchmarking AI Agent Harnesses for Reliability and Failure Attribution

The paper introduces TraceTrial-Bench, a benchmark that evaluates the reliability of AI agent harnesses across seven components—context, execution, governance, lifecycle, observability, tools, and verification—using 95 expert-level tasks with executable or state-based ground truth. It employs a Trajectory Tracing module to record agent execution, a two-layer evaluation framework to separately assess agent performance and harness reliability, and a Multi-source Evidence Acquisition module to gather evidence for LLM-based evaluation. The benchmark’s analysis of 3,420 runs across four harnesses and nine models shows harness scores between 59.8% and 72.0%, with 31.1% of runs succeeding despite failing at least one harness criterion, highlighting the need for auditable and reproducible harness evaluation.

By Zhiyong Cao, Hengjie Zhu, Yongxin Wang, Xuchen Li, Jintao Li, Shuxia Liu, Dunqiang Liu, Bo Wang, Kairan Zhang, Cong Lin, JiangDi Yu, Siyuan Liu, Sijia Yao, Yi-Liang Wei, Lingxu Ma, Aijia Wang, Anxin Chen, Zixian Wei, Xiang Ruan, Ruizhi Xu, Ruqian Shi, Baihua Xiao, Sicheng Zhou