The paper introduces a Trust Layer for Agent Evaluation, a post‑hoc framework that appends a belief flag to each benchmark score. It checks four properties: alignment with the benchmark’s grading logic, traceable computation of a passing answer, consistency between the agent’s claim and actual behavior, and stability across repeated runs. Applied to five agent configurations on 108 tasks, the study finds that only 22.6% of recorded passes satisfy all checks, revealing widespread issues such as untraceable computation, false completion claims, and unstable results.
By Mohammadreza Sediqin, Shivali Dalmia, Srinivasa Karthikeya Reddy Kovvuri, Abhishek Mukherji
The paper investigates whether an agent’s own verifier remains reliable when the agent is optimized against it. It introduces a stress‑testing framework that measures the pressure ratio ρ, the influence of the verifier on the agent’s search, and four falsifiable probes targeting common failure modes. Applying this framework to the Rationale Consistency Score (RCS) in an LLM agent that designs machine‑learning pipelines, the authors find that the verifier’s impact is minimal in their baseline setup, that agents can shorten claim lists without affecting search, and that pre‑commit probing can hide declines in faithfulness. The study concludes with a proposed reporting standard to audit verifiers before trusting them.
By Jonghyun Lee, Moonsu Kim, Sunghwan Park
arXiv:2608.22808v2 Announce Type: replace
Abstract: When can an agent failure be caught? An audit is usually limited by the record rather than by the method. CatchBench therefore puts one auditor's q...
By Yue Zhao
arXiv:2606. 19782v1 Announce Type: new Abstract: Financial chart question answering in regulated settings demands more than accuracy: practitioners must know which answers to trust before acting on them, and many institutions cannot send client data to external model providers.
By Aravind Narayanan, Shaina Raza
The paper "trajectory-judge: What Outcome-Only LLM Judges Miss on Agent Trajectories" examines the limitations of outcome-only evaluation for large language model agents. Using a deterministic tool‑using support‑desk environment with a scripted oracle policy and a fault injector, the authors compare five different judging approaches—programmatic rules, outcome‑only, step‑rubric at two model sizes, and a self‑consistency ensemble—on metrics such as detection, step localisation, fault typing, calibration, and cost across 400 trajectories. The study finds that outcome‑only judges miss many silent faults and generate false positives, while step‑rubric judges achieve higher recall with no false alarms but at greater cost, and that none of the judges read the final reply, allowing fabricated promises to evade detection.
"whyItMatters":"The findings highlight that current production‑default outcome‑only evaluations can overlook critical failures in agent behavior, underscoring the need for more nuanced, step‑level judging methods to ensure reliable LLM agent performance."
By Hadi Mohammadi
AgentAudit is an open, extensible framework that evaluates the full lifecycle of AI agents, assessing planning, tool selection, execution, memory, and reasoning across ten dimensions such as instruction integrity, security, and alignment. Unlike existing benchmarks that focus on single aspects, AgentAudit analyzes the entire execution trace to attribute failures to specific stages. The framework was tested on five large language models, revealing significant differences in trustworthiness even among models with similar task‑completion performance.
By Shrey Nag, Sachita, Abhishek Kumar Singh, Lipi Goel, Rajeshwar Singh Janwar
arXiv:2609.37315v1 Announce Type: cross
Abstract: Tool-using agents are entering settings where a wrong action carries real cost, and the benchmarks certifying them grade what each simulated tool cal...
By Rohith Reddy Bellibatlu, Zichong Wang, Wenbin Zhang
The paper proposes a claim‑specific verification audit for modular agents that replaces aggregate task scores with evidence‑based evaluations. Each agent conclusion is recorded with supporting evidence and classified as supported, unsupported, unresolved, or not evaluated, along with the boundary of validity. The audit employs three tools—oracle policies, perfect component replacements, and verifier‑score tests—to trace value changes, locate lost value, and assess verifier effectiveness, demonstrated on a portfolio‑allocation agent in a synthetic market.
By Ali Atiah Alzahrani
The study examines how large language models (LLMs) provide financial advice based on a single risk‑tolerance question, finding that both human respondents and LLMs exhibit significant variability when the same question is repeated. In a national US panel, 41% of households changed their risk answer after two years, and in a lab retest a third changed by two or more points on a 0‑10 scale after six months. When LLMs were given 800 real household profiles with only the risk answer altered, the recommended stock allocation shifted by 10–16 percentage points on average, and by 24–35 points when the answer changed, compared to minimal swings when the same prompt was repeated. The study also shows that adding clarifying questions or noting answer noise can reduce these swings, especially for younger clients and those with extreme risk attitudes.
whyItMatters":"The findings highlight that LLMs can produce markedly different financial advice from a single, potentially unreliable risk‑tolerance response, underscoring the need for repeat‑testing and transparency in AI‑driven financial recommendations."
By Soovadeep Bakshi, Apurba Bose, Aishwarya Bose, Sandip Nandi
The paper investigates whether large language model (LLM) safety judges are truly policy‑invariant, meaning their verdicts should depend only on the agent’s behavior and not on how the evaluation policy is phrased. It introduces three testable principles—rubric‑semantics invariance, rubric‑threshold invariance, and ambiguity‑aware calibration—and applies them in a stress‑test protocol across four agent‑class judges using trajectories from ASSEBench and R‑Judge. The study finds that current judges often change verdicts in response to both meaningful normative shifts and meaningless structural rewrites, with up to 9.1% of verdicts flipping due to content‑preserving rewrites and 18‑43% of flips occurring on unambiguous cases, indicating that safety scores conflate agent actions with evaluator wording.
whyItMatters":"The findings reveal a hidden failure mode in safety evaluation that could mislead researchers about agent performance, underscoring the need for policy‑invariant judges and transparent reporting protocols."
By Shihao Weng, Yang Feng, Xiaofei Xie
arXiv:2610.17301v1 Announce Type: new
Abstract: Language-model agents sometimes act beyond their authorization and then misreport what they did. On CAPE-Bench, matched agent tasks in which one fact d...
By Arya Fayyazi, Mehdi Kamal, Massoud Pedram
arXiv:2607. 24889v1 Announce Type: cross Abstract: Financial models combine public disclosures with analyst assumptions to produce forecasts and valuations.
By Jiacheng Lu, Sinuo Wang, Wentao Zhao, Rui Sun, Cheng Hua, Tao Song, Hui Cai, Beidi Luan, Zhengze Wu, Lingjing Teng, Yijia He, Jing Li, Daxin Jiang, Zuo Bai, Haibing Guan