AgentHallu: Benchmarking Automated Hallucination Attribution of LLM-based Agents

TL;DR

This paper introduces AgentHallu, a benchmark for identifying and explaining the specific steps causing hallucinations in multi-step reasoning of LLM-based agents, highlighting the challenge and need for improved reliability.

Contribution

It presents a new task, dataset, and evaluation framework for hallucination attribution in multi-step LLM agents, filling a critical gap in understanding and diagnosing hallucinations.

Findings

Top models achieve only 41.1% accuracy in step localization.

Tool-use hallucinations are the most challenging to identify.

The task remains difficult even for state-of-the-art models.

Abstract

As LLM-based agents operate over sequential multi-step reasoning, hallucinations arising at intermediate steps risk propagating along the trajectory, thus degrading overall reliability. Unlike hallucination detection in single-turn responses, diagnosing hallucinations in multi-step workflows requires identifying which step causes the initial divergence. To fill this gap, we propose a new research task, automated hallucination attribution of LLM-based agents, aiming to identify the step responsible for the hallucination and explain why. To support this task, we introduce AgentHallu, a comprehensive benchmark with: (1) 693 high-quality trajectories spanning 7 agent frameworks and 5 domains, (2) a hallucination taxonomy organized into 5 categories (Planning, Retrieval, Reasoning, Human-Interaction, and Tool-Use) and 14 sub-categories, and (3) multi-level annotations curated by humans, covering binary labels, hallucination-responsible steps, and causal explanations. We evaluate 13 leading models, and results show the task is challenging even for top-tier models (like GPT-5, Gemini-2.5-Pro). The best-performing model achieves only 41.1\% step localization accuracy, where tool-use hallucinations are the most challenging at just 11.6\%. We believe AgentHallu will catalyze future research into developing robust, transparent, and reliable agentic systems.