Answering the Unanswerable Is to Err Knowingly: Analyzing and Mitigating Abstention Failures in Large Reasoning Models
On the map: LLM reasoning → Test-time compute and efficient thinking
Abstract (arXiv)
Large reasoning models (LRMs) have shown remarkable progress on complex reasoning tasks. However, some questions posed to LRMs are inherently unanswerable, such as math problems lacking sufficient conditions. We find that LRMs continually fail to provide appropriate abstentions when confronted with these unanswerable questions. In this paper, we systematically analyze, investigate, and resolve this issue for trustworthy AI. We first conduct a detailed analysis of the distinct response behaviors of LRMs when facing unanswerable questions. Then, we show that LRMs possess sufficient cognitive capabilities to recognize the flaws in these questions. However, they fail to exhibit appropriate abstention behavior, revealing a misalignment between their internal cognition and external response. Finally, to resolve this issue, we propose a lightweight, two-stage method that combines cognitive monitoring with inference-time intervention. Experimental results demonstrate that our method significantly improves the abstention rate while maintaining the overall reasoning performance.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 16 of its 33 references that are in the map. First posted on arXiv 26 Aug 2025.
Builds on
16 of 33 references.
Closest ancestors · 6
Same direction, newest first
- 2025The First Impression Problem: Internal Bias Triggers Overthinking in Reasoning Models22 May
- 2025When Can Large Reasoning Models Save Thinking? Mechanistic Analysis of Behavioral Divergence in Reasoning21 May
- 2025Dynamic Early Exit in Reasoning Models22 Apr
- 2025SEAL: Steerable Reasoning Calibration of Large Language Models for Free7 Apr
- 2025Reward-Guided Speculative Decoding for Efficient LLM Reasoning31 Jan
- 2025Thoughts Are All Over the Place: On the Underthinking of o1-Like LLMs30 Jan
Other references · 10
- 2021Training Verifiers to Solve Math Word Problemscited by 10,109
- 2025Qwen3 Technical Reportcited by 8,820
- 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
- 2024OpenAI o1 System Cardcited by 1,244
- 2023The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasetscited by 712
- 2025Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Modelscited by 293
- 2025AbstentionBench: Reasoning LLMs Fail on Unanswerable Questionscited by 145
- 2024Uncertainty-Based Abstention in LLMs Improves Safety and Reduces Hallucinationscited by 54
- 2025The Hallucination Tax of Reinforcement Finetuningcited by 40
- 2025TreeCut: A Synthetic Unanswerable Math Word Problem Dataset for LLM Hallucination Evaluationcited by 12
Built on by
Citing papers per year: 2025: 1; 2026: 2.
3 papers.
Similar, not cited
Close in wording, no citation either way
- 2025Stop Before You Fail: Operational Capability Boundaries for Mitigating Unproductive Reasoning in Large Reasoning Models
- 2026Bridging the Detection-to-Abstention Gap in Reasoning Models under Insufficient Information
- 2026Rewarding Efficient Reasoning Improves Abstention on Underspecified Tasks in Reasoning Models
- 2025Beyond Solving Math Quiz: Evaluating the Ability of Large Reasoning Models to Ask for Information
- 2025Exploring the System 1 Thinking Capability of Large Reasoning Models
- 2025The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
- 2025Trade-offs in Large Reasoning Models: An Empirical Analysis of Deliberative and Adaptive Reasoning over Foundational Capabilities
- 2026REIN: Bridging the Gap between Reasoning and Reliability via Reflection and Abstention Alignment