ThinkEdit: Interpretable Weight Editing to Mitigate Overly Short Thinking in Reasoning Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies have shown that Large Language Models (LLMs) augmented with chain-of-thought (CoT) reasoning demonstrate impressive problem-solving abilities. |
| Approach: | They propose a weight-editing approach to reduce overly short reasoning by steering the model along a linear direction in the representation space. |
| Outcome: | The proposed model reduces overly short reasoning and yields significant accuracy gains on multiple math benchmarks. |
Similar Papers
Too Long, Do Re-weighting for Efficient LLM Reasoning Compression (2026.acl-long)
Copied to clipboard
Zhong-Zhi Li, Xiao Liang, Zihao Tang, Lei Ji, Peijie Wang, Haotian Xu, Xing W, Haizhen Huang, Weiwei Deng, Yeyun Gong, Zhijiang Guo, Xiao Liu, Fei Yin, Cheng-Lin Liu
| Challenge: | Large Language Models (LLMs) have recently achieved remarkable progress on complex reasoning tasks by leveraging extended Chain-of-Thought (CoT) techniques. |
| Approach: | They propose a method that uses Extended Chain-of-Thought (EFT) to reduce the number of output tokens by nearly 40% while maintaining the accuracy of the reasoning. |
| Outcome: | The proposed method reduces the number of output tokens by nearly 40% while maintaining the accuracy of the reasoning. |
Feature Extraction and Steering for Enhanced Chain-of-Thought Reasoning in Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) can solve reasoning and mathematical problems using the Chain-of-Thought technique, but require costly and long CoT data and fine-tuning. |
| Approach: | They propose a method that uses Sparse Autoencoders to extract interpretable features from vanilla CoT and use them to steer the LLM's internal states. |
| Outcome: | The proposed method uses Sparse Autoencoders (SAEs) to extract interpretable features from vanilla CoT and steer the LLM's internal states during generation. |
The Impact of Reasoning Step Length on Large Language Models (2024.findings-acl)
Copied to clipboard
| Challenge: | Long reasoning steps in LLMs improve reasoning abilities, but the correlation between their effectiveness and the length of reasoning steps remains largely unknown. |
| Approach: | They conducted experiments that expand and compress the rationale reasoning steps within CoT demonstrations while keeping all other factors constant. |
| Outcome: | The results show that lengthening the reasoning steps in prompts significantly enhances LLMs’ reasoning abilities across multiple datasets. |
AutoL2S: Auto Long-Short Reasoning for Efficient Large Language Models (2026.findings-acl)
Copied to clipboard
Feng Luo, Yu-Neng Chuang, Guanchu Wang, Hoang Anh Duy Le, Shaochen Zhong, Hongyi Liu, Jiayi Yuan, Yang Sui, Vladimir Braverman, Vipin Chaudhary, Xia Hu
| Challenge: | Existing approaches to distilling large language models (LLMs) are inefficient and generate excessively long chain-of-thought reasoning even for inputs that admit concise solutions. |
| Approach: | They propose a distillation framework that empowers non-reasoning LLMs to think only when necessary. |
| Outcome: | The proposed framework reduces reasoning length up to 71% with minimal accuracy loss while preserving accuracy. |
Correct, Concise and Complete: Multi-stage Training For Adaptive Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) increase test-time computation, often in the form of chain-of-thought (CoT) however, reasoning traces can become unnecessarily long, increasing computation costs without improving accuracy and sometimes even degrading performance. |
| Approach: | They propose a multi-stage efficient reasoning method that combines supervised fine-tuning with reinforcement learning using an adaptive length penalty. |
| Outcome: | The proposed method reduces response length by an average of 28% for 8B models and 40% for 32B models while incurring only minor performance drops of 1.6 and 2.5 points, respectively. |
Do LLMs Really Need 10+ Thoughts for “Find the Time 1000 Days Later”? Towards Structural Understanding of LLM Overthinking (2026.acl-long)
Copied to clipboard
Xinliang Frederick Zhang, Anhad Mohananey, Alexandra Chronopoulou, Pinelopi Papalampidi, Somit Gupta, Tsendsuren Munkhdalai, Lu Wang, Shyam Upadhyay
| Challenge: | Existing studies on LLMs' thought processes are limited to superficial, profiling-based observations, failing to delve into their inner workings. |
| Approach: | They propose a utility-based definition of overthinking that moves beyond length-based metrics and provides a more insightful understanding of LLMs' thought progression. |
| Outcome: | The proposed model decomposes the LLM thought process into minimally complete sub-thoughts and identifies common thinking patterns for topically similar queries. |
Over-Reasoning and Redundant Calculation of Large Language Models (2024.eacl-short)
Copied to clipboard
| Challenge: | Large language models (LLMs) can solve problems step-by-step, but it is unclear whether they know when to use CoT and whether they are always necessary. |
| Approach: | They propose to use LLMs to generate redundant calculations and reasoning on a manually constructed math QA dataset, GSM8K-Zero. |
| Outcome: | The proposed model generates redundant calculations and reasoning on a manually constructed math QA dataset, but it is unclear whether it is necessary to use CoT reasoning. |
Improve Vision Language Model Chain-of-thought Reasoning (2025.acl-long)
Copied to clipboard
Ruohong Zhang, Bowen Zhang, Yanghao Li, Haotian Zhang, Zhiqing Sun, Zhe Gan, Yinfei Yang, Ruoming Pang, Yiming Yang
| Challenge: | Current training recipes often rely on datasets dominated by short annotations with limited rationales, hindering the models' ability to generalize to tasks requiring comprehensive reasoning. |
| Approach: | They propose a two-stage post-training strategy that augments short answers with CoT reasoning generated by GPT-4o, enhancing the VLM's CoT capabilities through fine-tuning. |
| Outcome: | The proposed strategy enhances the model's CoT capabilities through fine-tuning and reinforcement learning. |
Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketching (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in large language models (LLMs) have enabled strong reasoning capabilities through Chain-of-Thought (CoT) prompting. |
| Approach: | They propose a framework that integrates cognitively inspired reasoning paradigms with linguistic constraints to reduce token usage while preserving reasoning accuracy. |
| Outcome: | The proposed framework reduces token usage while preserving reasoning accuracy across 18 reasoning datasets across multiple domains, languages, and modalities. |
DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models (2025.emnlp-industry)
Copied to clipboard
Yi Shen, Jian Zhang, Jieyun Huang, Shuming Shi, Wenjing Zhang, Jiangze Yan, Ning Wang, Kai Wang, Zhaoxiang Liu, Shiguo Lian
| Challenge: | Recent advances in slow-thinking reasoning models have shown exceptional performance in complex reasoning tasks. |
| Approach: | They propose a framework that enables models to automatically adjust Chain-of-Thought (CoT) length based on problem difficulty. |
| Outcome: | The proposed framework penalizes inefficiency on simple problems while incentivizing deep reasoning for complex ones. |