Beyond Final Answers: Auditing Trajectory-Level Hallucinations in Multi-Agent Industrial Workflows

• URL: http://arxiv.org/abs/2605.24219v2
• Date: Tue, 26 May 2026 16:24:40 GMT
• Title: Beyond Final Answers: Auditing Trajectory-Level Hallucinations in Multi-Agent Industrial Workflows
• Abstract summary: Trajel is a dataset and evaluation framework for auditing trajectory-level hallucinations in multi-agent agents.<n>We benchmark supervised detection models at the subtask, trajectory, and long-context levels.<n>Trajectory-aware detection significantly outperforms standard post-hoc verification.
• Score: 1.6048497142263283
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Models (LLMs) are increasingly deployed as autonomous agents that reason, use tools, and act over multiple steps. Yet most hallucination benchmarks still evaluate only the final output, missing failures that originate in intermediate Thought-Action-Observation steps. We present Trajel, a dataset and evaluation framework for auditing trajectory-level hallucinations in multi-agent industrial workflows. Trajel introduces a five-type hallucination taxonomy (factual, referential, logical, procedural, and scope-based) over expert-annotated agent traces from AssetOpsBench. We benchmark supervised detection models at the subtask, trajectory, and long-context levels. Our results show that the most common failure modes are missed by existing benchmarks, that nearly half of hallucinated trajectories involve multiple types at once, and that automated detectors with high binary accuracy still misclassify the subtlest types. Trajectory-aware detection significantly outperforms standard post-hoc verification, making taxonomy-grounded evaluation necessary for safer agentic deployment.

Related papers

• A Multi-Agent Audit Framework for High-Stakes Reasoning: Evaluation and Interpretability in Clinical Mental Health Screening [0.8758828205574085]
  We propose a Multi-Agent Audit Framework that simulates a collaborative, multi-step verification process.<n>We empirically validate this architecture in the sensitive domain of clinical mental health screening.<n>Our framework decomposes the reasoning process into a Perception Agent, Knowledge Retrieval-Augmented Generation (RAG), Chain-of-Thought (CoT) clinical inference, and a critical Audit verification stage.
  arXiv  Detail & Related papers  (2026-06-19T05:51:03Z)
• ClawForge: Generating Executable Interactive Benchmarks for Command-Line Agents [59.626170560327274]
  We present textbfClawForge, a generator-backed benchmark framework for executable command-line categories under state conflict.<n>We instantiate this framework as the ClawForge-Bench (17 scenarios, 6 ability categories)
  arXiv  Detail & Related papers  (2026-05-13T21:34:08Z)
• How to Interpret Agent Behavior [56.59836196946289]
  We introduce ACT*ONOMY, a taxonomy for describing and analyzing agent behavior at runtime.<n>By providing a shared vocabulary, ACT*ONOMY helps researchers, agent designers, and end users interpret agent behavior more consistently.
  arXiv  Detail & Related papers  (2026-05-13T14:52:40Z)
• LiveFMBench: Unveiling the Power and Limits of Agentic Workflows in Specification Generation [75.05397479715576]
  Large language models (LLMs) and agents have shown promising progress, but their true capabilities and failure modes remain unclear.<n>We present the first systematic and contamination-aware study of LLM- and agent-based formal specification generation for C programs.
  arXiv  Detail & Related papers  (2026-05-02T11:31:33Z)
• AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using Agents [50.481033105867205]
  We introduce AgentProcessBench, the first benchmark dedicated to evaluating step-level effectiveness in realistic, tool-augmented trajectories.<n>The benchmark comprises 1,000 diverse trajectories and 8,509 human-labeled step annotations with 89.1% inter-annotator agreement.<n>It features a ternary labeling scheme to capture exploration and an error propagation rule to reduce labeling ambiguity.
  arXiv  Detail & Related papers  (2026-03-15T16:13:58Z)
• MAS-ProVe: Understanding the Process Verification of Multi-Agent Systems [59.20800753428596]
  We present MAS-ProVe, a systematic empirical study of process verification for multi-agent systems (MAS)<n>Our study spans three verification paradigms (LLM-as-a-Judge, reward models, and process reward models)<n>We find that process-level verification does not consistently improve performance and frequently exhibits high variance.
  arXiv  Detail & Related papers  (2026-02-03T03:30:36Z)
• The Bitter Lesson of Diffusion Language Models for Agentic Workflows: A Comprehensive Reality Check [54.08619694620588]
  We present a comprehensive evaluation of dLLMs across two distinct agentic paradigms: Embodied Agents and Tool-Calling Agents.<n>Our results on Agentboard and BFCL reveal a "bitter lesson": current dLLMs fail to serve as reliable agentic backbones.
  arXiv  Detail & Related papers  (2026-01-19T11:45:39Z)
• AgentIAD: Tool-Augmented Single-Agent for Industrial Anomaly Detection [9.208080780536504]
  AgentIAD is a tool-driven framework that enables multi-stage visual inspection.<n>AgentIAD achieves a new state-of-the-art 97.62% classification accuracy on MMAD.
  arXiv  Detail & Related papers  (2025-12-15T18:57:04Z)
• AgentCompass: Towards Reliable Evaluation of Agentic Workflows in Production [4.031479494871582]
  We present Agent, the first evaluation framework designed specifically for post-deployment monitoring and reasoning of agentic pipeline.<n>Agent achieves state-of-the-art results on key metrics, while uncovering critical issues missed in human annotations.
  arXiv  Detail & Related papers  (2025-09-18T05:59:04Z)
• TRAIL: Trace Reasoning and Agentic Issue Localization [5.025960714013197]
  This work articulates the need for robust and dynamic evaluation methods for agentic workflow traces.<n>We present a set of 148 large human-annotated traces (TRAIL) constructed using this taxonomy and grounded in established agentic benchmarks.<n>To ensure ecological validity, we curate traces from both single and multi-agent systems.
  arXiv  Detail & Related papers  (2025-05-13T14:55:31Z)

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.