{"slug": "rewarding-efficient-reasoning-improves-abstention-on-underspecified-tasks-in", "title": "Rewarding Efficient Reasoning Improves Abstention on Underspecified Tasks in Reasoning Models", "summary": "A new arXiv paper (2609.20846v1) reports that fine-tuning several 4B large reasoning models with a novel GRPO reward that encourages efficient reasoning about whether a task contains enough information improves abstention performance by 12.8% on average while shortening chains of thought by 44% on average. The authors compared large reasoning model behavior against a human study, finding that human reasoning effort on unanswerable tasks is upper-bounded by answerable tasks, whereas the models generate longer chains of thought on unanswerable than on answerable prompts. The reward is inspired by a resource-rational perspective on human cognition and retains the models' answering capabilities.", "body_md": "arXiv:2609.20846v1 Announce Type: new \nAbstract: While modern large reasoning models (LRMs) excel at providing correct answers in many tasks, we provide additional evidence for the observation that they often struggle with a critical capability: knowing when to abstain from answering. We analyze this gap by comparing LRM behavior to results from a human study, revealing that human reasoning effort on unanswerable tasks is upper-bounded by answerable tasks, whereas LRMs waste computational resources by generating longer Chains of Thought (CoTs) on unanswerable than on answerable prompts. To overcome this inefficiency, we take inspiration from a resource-rational perspective on human cognition and introduce a novel GRPO reward that encourages efficient reasoning about whether the task contains all the information needed to solve it. Fine-tuning several 4B LRMs with this reward leads to human-like abstention performance gains (+12.8% on average) while retaining answering capabilities and boosting the models' efficiency (44% shorter CoTs on average).", "url": "https://wpnews.pro/news/rewarding-efficient-reasoning-improves-abstention-on-underspecified-tasks-in", "canonical_source": "https://arxiv.org/abs/2609.20846", "published_at": "2026-09-21 04:00:00+00:00", "updated_at": "2026-09-21 04:23:46.491511+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "machine-learning", "artificial-intelligence"], "entities": ["arXiv", "GRPO"], "alternates": {"html": "https://wpnews.pro/news/rewarding-efficient-reasoning-improves-abstention-on-underspecified-tasks-in", "markdown": "https://wpnews.pro/news/rewarding-efficient-reasoning-improves-abstention-on-underspecified-tasks-in.md", "text": "https://wpnews.pro/news/rewarding-efficient-reasoning-improves-abstention-on-underspecified-tasks-in.txt", "jsonld": "https://wpnews.pro/news/rewarding-efficient-reasoning-improves-abstention-on-underspecified-tasks-in.jsonld"}}