paper-with-me

홈 › Papers

Bridging the Detection-to-Abstention Gap in Reasoning Models under Insufficient Information

2026-05-27 · Renjie Gu, Jiaxu Li, Yihao Wang, Yun Yue, Hansong Xiao, Yefei Chen, Yuan Wang, Chunxiao Guo, Pei Wei, Jinjie Gu, Yixin Cao arxiv

We highlight a failure mode of large reasoning models on questions with insufficient information: models may recognize that a problem is under-specified, yet still continue reasoning and produce unsupported final answers instead of abstaining. We formalize this mismatch as the detection-to-abstention gap, where detected insufficiency fails to translate into final abstention. This gap is especially concerning in high-risk domains such as medical AI, where answers based on incomplete evidence can be more harmful than refusal. To close this gap, we propose Judge-Then-Solve (JTS), a trajectory-level reasoning-control framework that trains models to make an explicit answerability commitment before solution generation. Rather than treating abstention as a final-answer style, JTS casts it as a control decision: the model either proceeds to solve or terminates early based on its answerability judgment. We instantiate this policy through supervised warm-up and missing-premise reinforcement learning with consistency and length-shaping rewards. Experiments on dense and MoE reasoning models show that JTS substantially improves reliable abstention across datasets and pushes Abstention@Detection (A@D) to near-saturation, indicating that models not only detect missing information but also act on that detection. By terminating unanswerable trajectories immediately after the answerability judgment, JTS reduces unnecessary reasoning and improves inference efficiency when continued deliberation would amplify unsupported assumptions. We also observe that missing-premise training can alter reasoning behavior on difficult but answerable problems, reducing unproductive self-reflection. These results suggest that abstention under insufficient information is a key form of reasoning control for deploying reasoning models safely and efficiently.

📄 PDF Abstract BibTeX arXiv:2605.28070

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

KnowGuard: Knowledge-Driven Abstention for Multi-Round Clinical Reasoning

2025-09-29 · Xilin Dang, Kexin Chen, Xiaorui Su, Ayush Noori 외 arxiv

In clinical practice, physicians refrain from making decisions when patient information is insufficient. This behavior, known as abstention, is a critical safety mechanism preventing potentially harmful misdiagnoses. Rec…

Twin Worlds: Equivariance-Based Abstention for Evidence-Grounded Reasoning

2026-08-28 · Vy Nguyen, Ziqi Xu, Jeffrey Chan, Estrid He 외 arxiv

Knowledge-intensive reasoning requires Large Language Models (LLMs) to ground answers in provided evidence. When evidence is insufficient, it is desirable that models abstain rather than confidently generating unsupporte…

AbstentionBench: Reasoning LLMs Fail on Unanswerable Questions

2025-06-10 · Polina Kirichenko, Mark Ibrahim, Kamalika Chaudhuri, Samuel J. Bell

For Large Language Models (LLMs) to be reliably deployed in both everyday and high-stakes domains, knowing when not to answer is equally critical as answering correctly. Real-world user queries, which can be underspecifi…

Math

Knowing When to Quit: A Principled Framework for Dynamic Abstention in LLM Reasoning

2026-04-20 · Hen Davidov, Nachshon Cohen, Oren Kalinsky, Yaron Fairstein 외 arxiv

LLMs utilizing chain-of-thought reasoning often waste substantial compute by producing long, incorrect responses. Abstention can mitigate this by withholding outputs unlikely to be correct. While most abstention methods …

Mathematical ReasoningReinforcement Learning

Rewarding Intellectual Humility Learning When Not To Answer In Large Language Models

2026-01-27 · Abha Jha, Akanksha Mahajan, Ashwath Vaithinathan Aravindan, Praveen Saravanan 외 arxiv

Large Language Models (LLMs) often produce hallucinated or unverifiable content, undermining their reliability in factual domains. This work investigates Reinforcement Learning with Verifiable Rewards (RLVR) as a trainin…

Reinforcement LearningQuestion Answering