Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories

• URL: http://arxiv.org/abs/2606.02060v2
• Date: Tue, 02 Jun 2026 10:30:22 GMT
• Title: Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories
• Abstract summary: Evaluation based on final answers shows whether an agent succeeds, but not which parts of the trajectory make the answer unreliable.<n>We collect 2,790 real trajectories from two agent frameworks, three backbone models, and three benchmarks, convert raw logs into semantic spans, and annotate harmful error spans through expert review.<n>We propose DRIFT, a claim-centric auditing framework that tracks agent claims, checks their support in trajectory evidence, and marks spans where unsupported or conflicting claims affect the answer path.
• Score: 51.22051230894794
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Deep-research agents solve tasks through long trajectories of search, tool use, evidence inspection, and answer synthesis. Evaluation based on final answers shows whether an agent succeeds, but not which parts of the trajectory make the answer unreliable. We study span-level error localization for deep-research agents. We collect 2,790 real trajectories from two agent frameworks, three backbone models, and three benchmarks, convert raw logs into semantic spans, and annotate harmful error spans through LLM-assisted expert review. From these annotations, we build TELBench, a 1,000-instance benchmark for identifying error spans among normal exploration, failed searches, tentative hypotheses, and harmless noise. We further propose DRIFT, a claim-centric auditing framework that tracks agent claims, checks their support in trajectory evidence, and marks spans where unsupported or conflicting claims affect the answer path. Experiments across model families and auditing frameworks show that DRIFT improves span-level error localization and first-error accuracy by up to 30 percentage points. Our work provides a process-level view of reliability in deep-research agents.

Related papers

• Who is the Agent to Blame? Localizing Faithfulness and Citation Mistakes in Agentic Deep Research [75.8770212601511]
  Deep research (DR) systems produce long-form cited reports by orchestrating multiple agents that search and synthesize information from the web.<n> Citations are the primary mechanism for evaluating the faithfulness of these reports, yet current DR systems exhibit poor citation recall.<n>We propose an evaluation method that pinpoints which agent introduced each error by locally testing agent invocations for faithfulness and verifiability relative to their own inputs.
  arXiv  Detail & Related papers  (2026-08-25T09:34:18Z)
• SearchAuditor: Auditing and Attributing Failures in Long-Horizon Search Agents [56.400332595452966]
  We introduce SearchAuditBench, a benchmark that evaluates whether auditors can localize, attribute, and repair search-agent failures.<n>SearchAuditBench comprises 1,243 failed trajectories, averaging 73.1 messages and 65.1K tokens, collected from eight open-weight models on five deep-search benchmarks.<n>We propose SearchAuditor, a multi-perspective auditing framework that effectively localizes, attributes, and repairs search-agent failures.
  arXiv  Detail & Related papers  (2026-08-05T09:09:00Z)
• Is Deep Research Reliable? Misleading Knowledge Induces False Conclusions [29.22067195388263]
  Deep Research agents conduct long-horizon investigations by iteratively planning, retrieving evidence, and generating reports.<n>It remains unclear whether they can resist apparently credible but factually false information introduced into these defenses.<n>We introduce MisKnow-Agent, a framework that constructs task-specific documents supporting manually audited false conclusions.
  arXiv  Detail & Related papers  (2026-07-23T03:28:08Z)
• RepoMirage: Probing Repository Context Reasoning in Code Agents with Perturbations [51.43574078961796]
  We introduce RepoMirage, a two-stage evaluation suite built on SWE-Bench Verified.<n>RepoMirage-Perturb applies three types of semantics-preserving repository-level perturbations, revealing a clear performance drop when correct solving requires broader context access.<n>We propose RepoAnchor, a structure-first prototype workflow that separates repository exploration from downstream problem solving.
  arXiv  Detail & Related papers  (2026-05-25T06:26:43Z)
• BLAgent: Agentic RAG for File-Level Bug Localization [2.2917707112773593]
  BLAgent is a novel agentic RAG framework for file-level bug localization.<n> BLAgent attains over 78% Top-1 accuracy with open-source models.<n> BLAgent improves end-to-end repair success by over 20%.
  arXiv  Detail & Related papers  (2026-05-18T07:20:13Z)
• Towards Self-Improving Error Diagnosis in Multi-Agent Systems [31.04613892300063]
  We introduce ErrorProbe, a self-improving framework for semantic failure attribution.<n>ErrorProbe identifies responsible agents and the originating error step.<n>It maintains a verified episodic memory that updates only when error patterns are confirmed by executable evidence.
  arXiv  Detail & Related papers  (2026-04-19T23:13:05Z)
• HiL-Bench (Human-in-Loop Benchmark): Do Agents Know When to Ask for Help? [32.54022440678003]
  coding agents solve complex tasks when given complete context but collapse when specifications are incomplete or ambiguous.<n>Current benchmarks are blind to this failure mode.<n>We present HiL-Bench to measure this selective escalation skill.
  arXiv  Detail & Related papers  (2026-04-10T15:21:44Z)
• To Search or Not to Search: Aligning the Decision Boundary of Deep Search Agents via Causal Intervention [61.82680155643223]
  We identify the root cause of misaligned decision boundaries, the threshold determining when accumulated information suffices to answer.<n>This causes over-search (redundant searching despite sufficient knowledge) and under-search (premature termination yielding incorrect answers.<n>We propose a comprehensive framework comprising two key components. First, we introduce causal intervention-based diagnosis that identifies boundary errors.<n>Second, we develop Decision Boundary Alignment for Deep Search agents (DAS)<n>Our DAS method effectively calibrates these boundaries, mitigating both over-search and under-search to achieve substantial gains in accuracy and efficiency.
  arXiv  Detail & Related papers  (2026-02-03T09:29:06Z)
• Where LLM Agents Fail and How They can Learn From Failures [62.196870049524364]
  Large Language Model (LLM) agents have shown promise in solving complex, multi-step tasks.<n>They amplify vulnerability to cascading failures, where a single root-cause error propagates through subsequent decisions.<n>Current systems lack a framework that can comprehensively understand agent error in a modular and systemic way.<n>We introduce the AgentErrorTaxonomy, a modular classification of failure modes spanning memory, reflection, planning, action, and system-level operations.
  arXiv  Detail & Related papers  (2025-09-29T18:20:27Z)
• Abduct, Act, Predict: Scaffolding Causal Inference for Automated Failure Attribution in Multi-Agent Systems [20.846301581161978]
  Failure attribution in multi-agent systems is a critical yet unsolved challenge.<n>Current methods treat this as a pattern recognition task over long conversation logs.<n>A2P Scaffolding transforms failure attribution from pattern recognition into a structured causal inference task.
  arXiv  Detail & Related papers  (2025-09-12T16:51:15Z)
• Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems [50.29939179830491]
  Failure attribution in LLM multi-agent systems remains underexplored and labor-intensive.<n>We develop and evaluate three automated failure attribution methods, summarizing their corresponding pros and cons.<n>The best method achieves 53.5% accuracy in identifying failure-responsible agents but only 14.2% in pinpointing failure steps.
  arXiv  Detail & Related papers  (2025-04-30T23:09:44Z)

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.