English
Related papers

Related papers: Entropy-Guided Reasoning Compression

200 papers

Dimensionality reduction, a form of compression, can simplify representations of information to increase efficiency and reveal general patterns. Yet, this simplification also forfeits information, thereby reducing representational capacity.…

Large reasoning models (LRMs) excel on complex problems but face a critical barrier to efficiency: reinforcement learning (RL) training requires long rollouts for outcome-based rewards, where autoregressive decoding dominates time and…

Machine Learning · Computer Science 2026-02-20 Zeliang Zhang , Xiaodong Liu , Hao Cheng , Hao Sun , Chenliang Xu , Jianfeng Gao

Modern reasoning models, such as OpenAI's o1 and DeepSeek-R1, exhibit impressive problem-solving capabilities but suffer from critical inefficiencies: high inference latency, excessive computational resource consumption, and a tendency…

Computation and Language · Computer Science 2025-08-05 Hang Yuan , Bin Yu , Haotian Li , Shijun Yang , Christina Dan Wang , Zhou Yu , Xueyin Xu , Weizhen Qi , Kai Chen

Large language models have shown remarkable performance across a wide range of language tasks, owing to their exceptional capabilities in context modeling. The most commonly used method of context modeling is full self-attention, as seen in…

Computation and Language · Computer Science 2025-06-26 Zhisong Zhang , Yan Wang , Xinting Huang , Tianqing Fang , Hongming Zhang , Chenlong Deng , Shuaiyi Li , Dong Yu

Reinforcement Learning (RL)-based post-training has significantly advanced the complex reasoning capabilities of language models, fostering sophisticated self-reflection processes. However, this ``slow thinking'' paradigm presents a…

Machine Learning · Computer Science 2025-06-24 Xu Wan , Wei Wang , Wenyue Xu , Wotao Yin , Jie Song , Mingyang Sun

Latent reasoning via continuous chain-of-thoughts (Latent CoT) has emerged as a promising alternative to discrete CoT reasoning. Operating in continuous space increases expressivity and has been hypothesized to enable superposition: the…

Computation and Language · Computer Science 2026-04-09 Michael Rizvi-Martel , Guillaume Rabusseau , Marius Mosbach

An effective way to scale up test-time compute of large language models is to sample multiple responses and then select the best one, as in Grok Heavy and Gemini Deep Think. Existing selection methods often rely on external reward models,…

Machine Learning · Computer Science 2026-05-04 Wenshuo Zhao , Qi Zhu , Xingshan Zeng , Fei Mi , Lifeng Shang , Yi R. , Fung

Test-time scaling has proven effective in further enhancing the performance of pretrained Large Language Models (LLMs). However, mainstream post-training methods (i.e., reinforcement learning (RL) with chain-of-thought (CoT) reasoning)…

Machine Learning · Computer Science 2025-08-19 Yuyang Xu , Yi Cheng , Haochao Ying , Zhuoyun Du , Renjun Hu , Xing Shi , Wei Lin , Jian Wu

Reasoning is a key component of language understanding in Large Language Models. While Chain-of-Thought prompting enhances performance via explicit intermediate steps, it suffers from sufficient token overhead and a fixed reasoning…

Computation and Language · Computer Science 2025-11-18 Xinyuan Wang , Dongjie Wang , Wangyang Ying , Haoyue Bai , Nanxu Gong , Sixun Dong , Kunpeng Liu , Yanjie Fu

Large reasoning models (LRMs) like OpenAI o1 and DeepSeek-R1 achieve high accuracy on complex tasks by adopting long chain-of-thought (CoT) reasoning paths. However, the inherent verbosity of these processes frequently results in redundancy…

Computation and Language · Computer Science 2026-03-10 Chenzhi Hu , Qinzhe Hu , Yuhang Xu , Junyi Chen , Ruijie Wang , Shengzhong Liu , Jianxin Li , Fan Wu , Guihai Chen

In "Compress or Route?" (Johnson, 2026), we found that code generation tolerates aggressive prompt compression (r >= 0.6) while chain-of-thought reasoning degrades gradually. That study was limited to HumanEval (164 problems), left the…

Computation and Language · Computer Science 2026-02-19 Warren Johnson

Reasoning models often exhibit overthinking, characterized by redundant reasoning steps. We identify \emph{internal bias} elicited by the input question as a key trigger of such behavior. Upon encountering a problem, the model immediately…

Artificial Intelligence · Computer Science 2026-03-03 Renfei Dang , Zhening Li , Shujian Huang , Jiajun Chen

The method of maximum entropy has been very successful but there are cases where it has either failed or led to paradoxes that have cast doubt on its general legitimacy. My more optimistic assessment is that such failures and paradoxes…

Data Analysis, Statistics and Probability · Physics 2015-06-12 Ariel Caticha

Long chain-of-thought (CoT) significantly enhances the reasoning capabilities of large language models (LLMs). However, extensive reasoning traces lead to inefficiencies and increased time-to-first-token (TTFT). We propose a training…

Computation and Language · Computer Science 2026-01-08 Roy Xie , David Qiu , Deepak Gopinath , Dong Lin , Yanchao Sun , Chong Wang , Saloni Potdar , Bhuwan Dhingra

We present the surprising finding that a language model's reasoning capabilities can be improved by training on synthetic datasets of chain-of-thought (CoT) traces from more capable models, even when all of those traces lead to an incorrect…

Artificial Intelligence · Computer Science 2026-01-26 Abhranil Chandra , Ayush Agrawal , Arian Hosseini , Sebastian Fischmeister , Rishabh Agarwal , Navin Goyal , Aaron Courville

Entropy-based confidence signals are increasingly leveraged to improve reasoning in large language models (LLMs), yet existing approaches treat confidence as a static quantity -- typically aggregated over tokens. We show that the…

Machine Learning · Computer Science 2026-03-09 Chenghua Zhu , Siyan Wu , Xiangkang Zeng , Zishan Xu , Zhaolu Kang , Yifu Guo , Yuquan Lu , Junduan Huang , Guojing Zhou

Large reasoning models with reasoning capabilities achieve state-of-the-art performance on complex tasks, but their robustness under multi-turn adversarial pressure remains underexplored. We evaluate nine frontier reasoning models under…

Artificial Intelligence · Computer Science 2026-03-13 Yubo Li , Ramayya Krishnan , Rema Padman

Reinforcement learning from verifiable rewards has significantly advanced the reasoning capabilities of large language models. However, Group Relative Policy Optimization (GRPO) typically assigns a uniform, sequence-level advantage to all…

Machine Learning · Computer Science 2026-04-06 Song Yu , Li Li , Wenwen Zhao , Zhisheng Yang

Language models often generate long chain-of-thought traces, but it remains unclear how much of this reasoning is necessary for preserving the final prediction. We study this through the lens of overcomplete reasoning traces: generated…

Artificial Intelligence · Computer Science 2026-05-15 Sanjoy Chowdhury , Dinesh Manocha

In this work, we investigate how a model's tendency to broadly integrate its parametric knowledge evolves throughout pretraining, and how this behavior affects overall performance, particularly in terms of knowledge acquisition and…

Computation and Language · Computer Science 2025-03-13 Jiyeon Kim , Hyunji Lee , Hyowon Cho , Joel Jang , Hyeonbin Hwang , Seungpil Won , Youbin Ahn , Dohaeng Lee , Minjoon Seo