Pith. sign in

REVIEW 8 cited by

Revisiting Overthinking in Long Chain-of-Thought from the Perspective of Self-Doubt

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2505.23480 v1 pith:QBAYXLYY submitted 2025-05-29 cs.CL

Revisiting Overthinking in Long Chain-of-Thought from the Perspective of Self-Doubt

classification cs.CL
keywords overthinkingreasoningself-doubtlonganswermethodanalysischain-of-thought
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Reasoning Large Language Models (RLLMs) have demonstrated impressive performance on complex tasks, largely due to the adoption of Long Chain-of-Thought (Long CoT) reasoning. However, they often exhibit overthinking -- performing unnecessary reasoning steps even after arriving at the correct answer. Prior work has largely focused on qualitative analyses of overthinking through sample-based observations of long CoTs. In contrast, we present a quantitative analysis of overthinking from the perspective of self-doubt, characterized by excessive token usage devoted to re-verifying already-correct answer. We find that self-doubt significantly contributes to overthinking. In response, we introduce a simple and effective prompting method to reduce the model's over-reliance on input questions, thereby avoiding self-doubt. Specifically, we first prompt the model to question the validity of the input question, and then respond concisely based on the outcome of that evaluation. Experiments on three mathematical reasoning tasks and four datasets with missing premises demonstrate that our method substantially reduces answer length and yields significant improvements across nearly all datasets upon 4 widely-used RLLMs. Further analysis demonstrates that our method effectively minimizes the number of reasoning steps and reduces self-doubt.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Know When to Stop: Segment-Level Credit Assignment for Reducing Overthinking

    cs.CL 2026-07 unverdicted novelty 7.0

    DASH assigns segment-level credit in reasoning traces using drift toward ground-truth answers, yielding 50.8% accuracy on AIME25 versus 45.4% for GRPO while reducing overthinking behaviors.

  2. Diagnosing and Mitigating Thinking Collapse in On-Policy Self-Distillation

    cs.CL 2026-07 conditional novelty 6.5

    Thinking Collapse in reasoning OPSD is driven by teacher gradients at high-entropy forks; AD-OPSD’s dual-perspective soft gate recovers thinking density and up to +4.1% average accuracy.

  3. A Temporal Reasoning Benchmarking Framework for LRMs via Difficulty-controlled and Dynamic Test Generation

    cs.SE 2026-07 conditional novelty 6.5

    TRACEBench shows LRM temporal accuracy falls almost linearly with a controllable Allen-algebra difficulty score, while ~28% of mid-sized model successes are spurious guesses.

  4. Dynamic Rollout Editing for Reducing Overthinking in RL-Trained Reasoning Models

    cs.CL 2026-06 unverdicted novelty 6.0

    Dynamic Rollout Editing reduces overthinking in RL-trained LLMs by editing post-answer continuations in successful rollouts and preferring the edited versions within GRPO groups.

  5. CLORE: Content-Level Optimization for Reasoning Efficiency

    cs.AI 2026-05 unverdicted novelty 6.0

    CLORE augments correct on-policy rollouts by deleting repetitive and irrelevant segments then optimizes with auxiliary DPO to improve accuracy-efficiency trade-off on math benchmarks.

  6. BLADE: Boundary-Expanded and Layer-Adaptive Dynamic Exit for Efficient LLM Reasoning

    cs.CL 2026-07 conditional novelty 5.0

    BLADE stops LLM reasoning early at sentence and self-doubt boundaries using a layer-adaptive probe, cutting tokens ~25% with near-matched accuracy on five math benchmarks.

  7. Better Starts, Better Ends: Bootstrapped Iterative Self-Reasoning Distillation for Compressed Reasoning

    cs.CL 2026-07 conditional novelty 5.0

    Warm-starting on-policy self-distillation with a correctness-filtered, prompt-switched SFT bootstrap improves compression and accuracy of reasoning traces across math benchmarks.

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

    cs.AI 2026-05 unverdicted novelty 5.0

    JTS trains reasoning models via supervised warm-up and missing-premise RL to make an explicit answerability commitment that triggers early termination on unanswerable inputs, raising Abstention@Detection near saturation.