DoubtLess: Training-Free Contrastive Decoding for Efficient Reasoning

Hugging Face Daily Papers Papers
chain-of-thought contrastive-decoding inference-efficiency reasoning-models training-free token-reduction

Summary

DoubtLess 提出一种免训练的对比解码方法，用一个更小的推理模型引导大模型输出更简洁的思维链，在七个竞赛数学、研究生科学与编程基准上将 Qwen3-32B 的 CoT 缩短 31% 且准确率不变，优于五个推理时基线。

Reasoning models achieve strong performance through long chains of thought (CoT), but often spend more tokens than needed to solve a problem. Existing ways to shorten the CoT require additional training for each model, add a fixed steering vector to the model's hidden states, or use decoding heuristics such as stopping early or banning words like "wait", which can reduce accuracy. We propose DoubtLess, a training-free decoding method in which a much smaller reasoning model steers the large one towards brevity. The small model reads the large model's CoT so far twice, once prompted to be brief (the brief thinker) and once prompted to be verbose (the verbose thinker), and the difference between its two logit vectors, scaled by a strength, is added to the large model's logits at every step of the CoT. On seven benchmarks of competition mathematics, graduate-level science and competitive programming, DoubtLess shortens the CoT of Qwen3-32B by 31% with unchanged accuracy, outperforming five inference-time baselines. We also show that DoubtLess reduces thinking tokens per turn by 20% in a self-debugging coding agent with unchanged task performance. It also generalizes to two other model families with unchanged accuracy. When combined with training-based compression, it further shortens the CoT of a model already trained for brevity. Ablations and behavioral analysis show that DoubtLess reduces repeated checking, with most token savings coming after the correct answer first appears in the CoT. Our code is available at https://github.com/leoluolol/DoubtLess.
Original Article
View Cached Full Text

Cached at: 10/14/26, 06:51 AM

Paper page - DoubtLess: Training-Free Contrastive Decoding for Efficient Reasoning

Source: https://huggingface.co/papers/2610.18862

Abstract

Reasoningmodelsachievestrongperformancethroughlongchainsofthought(CoT),butoftenspendmoretokensthanneededtosolveaproblem.ExistingwaystoshortentheCoTrequireadditionaltrainingforeachmodel,addafixedsteeringvectortothemodel’shiddenstates,orusedecodingheuristicssuchasstoppingearlyorbanningwordslike“wait“,whichcanreduceaccuracy.WeproposeDoubtLess,atraining-freedecodingmethodinwhichamuchsmallerreasoningmodelsteersthelargeonetowardsbrevity.Thesmallmodelreadsthelargemodel’sCoTsofartwice,oncepromptedtobebrief(thebriefthinker)andoncepromptedtobeverbose(theverbosethinker),andthedifferencebetweenitstwologitvectors,scaledbyastrength,isaddedtothelargemodel’slogitsateverystepoftheCoT.Onsevenbenchmarksofcompetitionmathematics,graduate-levelscienceandcompetitiveprogramming,DoubtLessshortenstheCoTofQwen3-32Bby31%withunchangedaccuracy,outperformingfiveinference-timebaselines.WealsoshowthatDoubtLessreducesthinkingtokensperturnby20%inaself-debuggingcodingagentwithunchangedtaskperformance.Italsogeneralizestotwoothermodelfamilieswithunchangedaccuracy.Whencombinedwithtraining-basedcompression,itfurthershortenstheCoTofamodelalreadytrainedforbrevity.AblationsandbehavioralanalysisshowthatDoubtLessreducesrepeatedchecking,withmosttokensavingscomingafterthecorrectanswerfirstappearsintheCoT.Ourcodeisavailableathttps://github.com/leoluolol/DoubtLess.

View arXiv pageView PDFGitHub0Add to collection

Get this paper in your agent:

hf papers read 2610\.18862

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.18862 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.18862 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.18862 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

REDaCT: Collapse-Free Unlearning in Large Reasoning Models

arXiv cs.LG

REDaCT提出一种面向大型推理模型的推理时无遗忘框架，通过二元泄漏检测器与纠正性嵌入优化，在彻底移除遗忘集信息的同时完全避免推理链崩溃，实现近乎完美的答案与CoT遗忘并显著保留推理能力。