Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning

arXiv:2607.29211v1 Announce Type: new Abstract: Large language models generate computationally expensive yet semantically void reasoning on beyond-capability tasks, creating risks where plausible-sounding but incorrect derivations mislead users. We characterize this extit{futile reasoning} phenomenon through systematic analysis, revealing universal capability overreach and systematic miscalibration between capability and behavior. The dominant failure mode is specious reasoning, which outputs look superficially valid but contain subtle errors, escalating with task difficulty. To address this, we introduce extbf{CaRL} (extbf{Ca}pability-extbf{a}ligned extbf{R}einforcement extbf{L}earning), which aligns model behavior with capability boundaries through reward shaping that incentivizes refusal over futile reasoning and hindsight refusal augmentation that converts failures into refusal supervision. Experiments demonstrate a substantial reduction in futile reasoning while preserving performance across task difficulties, effectively achieving capability-aligned behavior without sacrificing utility. footnote{https://github.com/icip-cas/Knowing-When-to-Quit}

Related

• LaTER: Efficient Test-Time Reasoning via Latent Exploration and Explicit Verification
• Pause or Fabricate? Training Language Models for Grounded Reasoning
• Self-Reflective Generation at Test Time
• FERA: Uncertainty-Aware Federated Reasoning for Large Language Models
• DiffAdapt: Difficulty-Adaptive Reasoning for Token-Efficient LLM Inference

Source: arXiv cs.CL | 2026-08-03