HiL-Bench (Human-in-Loop Benchmark): Do Agents Know When to Ask for Help?

• URL: http://arxiv.org/abs/2604.09408v2
• Date: Mon, 13 Apr 2026 04:15:58 GMT
• Title: HiL-Bench (Human-in-Loop Benchmark): Do Agents Know When to Ask for Help?
• Abstract summary: 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.
• Score: 32.54022440678003
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Frontier coding agents solve complex tasks when given complete context but collapse when specifications are incomplete or ambiguous. The bottleneck is not raw capability, but judgment: knowing when to act autonomously and when to ask for help. Current benchmarks are blind to this failure mode. They supply unambiguous detailed instructions and solely reward execution correctness, so an agent that makes a lucky guess for a missing requirement will score identically to one that would have asked to be certain. We present HiL-Bench (Human-in-the-Loop Benchmark) to measure this selective escalation skill. Each task contains human-validated blockers (missing information, ambiguous requests, contradictory information) that surface only through progressive exploration, not upfront inspection. Our core metric, Ask-F1, the harmonic mean of question precision and blocker recall, captures the tension between over-asking and silent guessing; its structure architecturally prevents gaming through question spam. Evaluation across SWE and text-to-SQL domains reveals a large universal judgment gap: no frontier model recovers more than a fraction of its full-information performance when deciding whether to ask. Failure analysis identifies three key help-seeking patterns: overconfident wrong beliefs with no gap detection; high uncertainty detection yet persistent errors; broad, imprecise escalation without self-correction. These consistent patterns confirm poor help-seeking is a model-level flaw, not task-specific. RL training on shaped Ask-F1 reward shows judgment is trainable: a 32B model improves both help-seeking quality and task pass rate, with gains that transfer across domains. The model does not learn domain-specific heuristics for when to ask; it learns to detect unresolvable uncertainty and act on it.

Related papers

• Ask, Condition or Abstain: Reinforcement Learning for Missing-Premise Reasoning [44.87905259927536]
  We present emphAsk-Condition-Abstain Reinforcement Learning (ACA-RL), a data-augmented RL framework for this setting.<n>Its reasoning-graph-guided pipeline converts well-posed problems into missing-premise training instances with localized gap annotations.<n>We also introduce the emphMissing-Premise Benchmark (MPB), a 274-instance human-verified benchmark spanning mathematical, logical, and real-world word problems.
  arXiv  Detail & Related papers  (2026-08-17T13:24:41Z)
• HopRefusalBench: Diagnosing Refusal Failures in Search-Augmented Agents for Multi-Hop Reasoning [14.484579581234327]
  We introduce HopRefusalBench, the first benchmark of refusal within multi-hop search.<n>It crosses three causes of unanswerability with root, middle, and terminal topologies.<n>We propose a final-outcome taxonomy spanning target-aware refusal, pseudo-refusal, hallucinated completion, and search-budget exhaustion.
  arXiv  Detail & Related papers  (2026-08-02T16:20:55Z)
• The Knowing-Saying Gap: When Probes See Errors that Confidence Misses [20.46211358172975]
  Linear probes detect corrupted context in language models with near-perfect accuracy, yet this does not translate into reliable failure prediction.<n>Across multi-hop arithmetic chains, probes that detect corruption turn out to be uninformative about final answer correctness.<n>As a real-time monitor, probe-based interventions are sharply model and error-type dependent.
  arXiv  Detail & Related papers  (2026-07-21T12:13:10Z)
• Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories [51.22051230894794]
  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.
  arXiv  Detail & Related papers  (2026-06-01T10:50:26Z)
• Decomposed Prompting Does Not Fix Knowledge Gaps, But Helps Models Say "I Don't Know" [47.930782177987446]
  Large language models often struggle to recognize their knowledge limits in closed-book question answering, leading to confident hallucinations.<n>We evaluate three task-equivalent prompting regimes: Direct, Assistive, and Incremental, across different model scales and multi-hop QA benchmarks.<n>Because factual knowledge is stable while hallucinations are agreement, cross-regime provides a precise signal of internal uncertainty.
  arXiv  Detail & Related papers  (2026-02-04T18:39:58Z)
• 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)
• Sample Smart, Not Hard: Correctness-First Decoding for Better Reasoning in LLMs [72.82403830490084]
  We argue that the decoding rule should be calibrated by correctness, not confidence alone.<n>We propose simple strategies that achieve this goal: Greedy-Threshold makes sampling greedy at very low confidence steps.<n>Together, our findings challenge prevailings about decoding under uncertainty and show gains across math and general reasoning benchmarks.
  arXiv  Detail & Related papers  (2025-10-07T14:46:12Z)
• 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)
• LoGU: Long-form Generation with Uncertainty Expressions [49.76417603761989]
  We introduce the task of Long-form Generation with Uncertainty(LoGU)<n>We identify two key challenges: Uncertainty Suppression and Uncertainty Misalignment.<n>Our framework adopts a divide-and-conquer strategy, refining uncertainty based on atomic claims.<n>Experiments on three long-form instruction following datasets show that our method significantly improves accuracy, reduces hallucinations, and maintains the comprehensiveness of responses.
  arXiv  Detail & Related papers  (2024-10-18T09:15:35Z)
• R-Tuning: Instructing Large Language Models to Say `I Don't Know' [66.11375475253007]
  Large language models (LLMs) have revolutionized numerous domains with their impressive performance but still face their challenges. Previous instruction tuning methods force the model to complete a sentence no matter whether the model knows the knowledge or not. We present a new approach called Refusal-Aware Instruction Tuning (R-Tuning) Experimental results demonstrate R-Tuning effectively improves a model's ability to answer known questions and refrain from answering unknown questions.
  arXiv  Detail & Related papers  (2023-11-16T08:45:44Z)

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.