mlsanityData to 14 Oct 2026

Settle: Learning When to Stop Reasoning

arXiv 2609.38997 ↗First posted 30 Sep 2026cs.CLcited by 0

New in LLM reasoning; no direction yet. LLM reasoning

Abstract (arXiv)

Reasoning models often continue generating after their answers have settled. Settle learns when to stop from answer stability in completed traces. It trains the existing end-of-reasoning token while keeping other predictions close to the base model, and requires only ordinary decoding at inference. On MATH-500 with Qwen3-4B, Settle reduces token count by 40% with a 0.5-percentage-point decrease in accuracy. It gains 6.16 percentage points over supervised fine-tuning on the same traces shortened at their first stable answer, at nearly identical token counts. Its stopping score predicts whether a correct answer will remain correct. Settle extends the accuracy-token-count Pareto frontier of the evaluated stopping methods.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 19 of its 26 references that are in the map. First posted on arXiv 30 Sep 2026.

Builds on

19 of 26 references.

References · 19

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  3. 2025Qwen3 Technical Reportcited by 8,820
  4. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  5. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  6. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  7. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
  8. 2022Solving Quantitative Reasoning Problems with Language Modelscited by 1,392
  9. 2025s1: Simple test-time scalingcited by 1,341
  10. 2025Dynamic Early Exit in Reasoning Modelscited by 219
  11. 2025ThinkPrune: Pruning Long Chain-of-Thought of LLMs via Reinforcement Learningcited by 192
  12. 2025Self-Training Elicits Concise Reasoning in Large Language Modelscited by 99
7 more
  1. 2025Answer Convergence as a Signal for Early Stopping in Reasoningcited by 40
  2. 2026Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reasoning Modelscited by 7
  3. 2025LYNX: Learning Dynamic Exits for Confidence-Controlled Reasoningcited by 6
  4. 2026TERMINATOR: Learning Optimal Exit Points for Early Stopping in Chain-of-Thought Reasoningcited by 6
  5. 2026ESTAR: Early-Stopping Token-Aware Reasoning For Efficient Inferencecited by 3
  6. 2026Rethinking RL for LLM Reasoning: It's Sparse Policy Selection, Not Capability Learningcited by 3
  7. 2026SABER: Stability-Aware Early Exit for LLM Reasoning via Adversarial Branch Probingcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Early Stopping for Large Reasoning Models via Confidence Dynamics
  2. 2026When Does Learning to Stop Help? A Cost-Aware Study of Early Exits in Reasoning Models
  3. 2026Large Language Models Decide Early and Explain Later
  4. 2026Statistical Early Stopping for Reasoning Models
  5. 2025Reasoning Models Know When They're Right: Probing Hidden States for Self-Verification
  6. 2026When Does a Language Model Commit? A Finite-Answer Theory of Pre-Verbalization Commitment
  7. 2025Entropy After </Think> for reasoning model early exiting
  8. 2026Reasoning Stabilization Point: A Training-Time Signal for Stable Evidence and Shortcut Reliance