VeriThinker: Learning to Verify Makes Reasoning Model Efficient

2025-05-23 · Zigeng Chen, Xinyin Ma, Gongfan Fang, Ruonan Yu, Xinchao Wang

Large Reasoning Models (LRMs) excel at complex tasks using Chain-of-Thought (CoT) reasoning. However, their tendency to overthinking leads to unnecessarily lengthy reasoning chains, dramatically increasing inference costs. To mitigate this issue, we introduce VeriThinker, a novel approach for CoT compression. Unlike conventional methods that fine-tune LRMs directly on the original reasoning task using synthetic concise CoT data, we innovatively fine-tune the model solely through an auxiliary verification task. By training LRMs to accurately verify the correctness of CoT solutions, the LRMs inherently become more discerning about the necessity of subsequent self-reflection steps, thereby effectively suppressing overthinking. Extensive experiments validate that VeriThinker substantially reduces reasoning chain lengths while maintaining or even slightly improving accuracy. When applied to DeepSeek-R1-Distill-Qwen-7B, our approach reduces reasoning tokens on MATH500 from 3790 to 2125 while improving accuracy by 0.8% (94.0% to 94.8%), and on AIME25, tokens decrease from 14321 to 10287 with a 2.1% accuracy gain (38.7% to 40.8%). Additionally, our experiments demonstrate that VeriThinker can also be zero-shot generalized to speculative reasoning. Code is available at https://github.com/czg1225/VeriThinker

📄 PDF Abstract BibTeX arXiv:2505.17941

Code (1)

czg1225/verithinker 공식 구현

Tasks

model

Learning to Self-Verify Makes Language Models Better Reasoners

2026-02-07 · Yuxin Chen, Yu Wang, Yi Zhang, Ziang Ye 외 arxiv

Recent large language models (LLMs) achieve strong performance in generating promising reasoning paths for complex tasks. However, despite powerful generation ability, LLMs remain weak at verifying their own answers, rev…

Reinforcement Learning

Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solvers

2024-08-12 · Zhenting Qi, Mingyuan Ma, Jiahang Xu, Li Lyna Zhang 외

This paper introduces rStar, a self-play mutual reasoning approach that significantly improves reasoning capabilities of small language models (SLMs) without fine-tuning or superior models. rStar decouples reasoning into…

GSM8K Math StrategyQA

Explainable Claim Verification via Knowledge-Grounded Reasoning with Large Language Models

2023-10-08 · Haoran Wang, Kai Shu

Claim verification plays a crucial role in combating misinformation. While existing works on claim verification have shown promising results, a crucial piece of the puzzle that remains unsolved is to understand how to ve…

Claim Verification Decision Making In-Context Learning Misinformation

RSAT: Structured Attribution Makes Small Language Models Faithful Table Reasoners

2026-04-30 · Jugal Gajjar, Kamalasankari Subramaniakuppusamy arxiv

When a language model answers a table question, users have no way to verify which cells informed which reasoning steps. We introduce RSAT, a method that trains small language models (SLMs, 1-8B) to produce step-by-step r…

Thinking Before Constraining: A Unified Decoding Framework for Large Language Models

2026-01-12 · Ngoc Trinh Hung Nguyen, Alonso Silva, Laith Zumot, Liubov Tupikina 외 arxiv

Natural generation allows Large Language Models (LLMs) to produce free-form responses with rich reasoning, yet the lack of structure makes outputs difficult to verify. Conversely, constrained decoding ensures standardize…