Awesome AI Agents › Multi-Agent Surveys

LightChen233/Awesome-Long-Chain-of-Thought-Reasoning

⭐ 647 added to this list on 2025-04-19 repository created 2025-03-13

The project "Awesome-Long-Chain-of-Thought-Reasoning" is a comprehensive repository dedicated to the latest advances in long chain-of-thought (Long CoT) reasoning for large language models (LLMs). It is closely associated with a survey paper titled "Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models," which provides an in-depth analysis of Long CoT techniques and their impact on enhancing reasoning capabilities in AI models. The repository serves as a valuable resource hub, compiling over 900 reviewed papers, research updates, and categorized content related to Long CoT reasoning. Long CoT reasoning is distinguished from traditional short chain-of-thought (Short CoT) by its ability to support deep reasoning, extensive exploration, and feasible reflection, enabling models to tackle more complex problems with greater coherence and efficiency. The repository offers a novel taxonomy to categorize current reasoning paradigms and explores key phenomena such as overthinking and test-time scaling, which are critical to understanding the behavior and optimization of reasoning models. The content is organized into several parts, including analysis and evaluation of Long CoT, deep reasoning formats and learning methods, feasible reflection techniques like feedback and refinement, and extensive exploration strategies. It also highlights future research directions such as agentic and embodied Long CoT, efficiency improvements, knowledge augmentation, multilingual and multimodal reasoning, and safety considerations. This project is highly relevant for researchers and practitioners interested in advancing the reasoning capabilities of large language models. It provides a structured overview of the state-of-the-art in Long CoT reasoning, facilitating further research and development in this emerging field of artificial intelligence.

https://github.com/LightChen233/Awesome-Long-Chain-of-Thought-Reasoning

agentartificial-intelligencechain-of-thoughtdeep-reasoningdeepseek-r1efficiencyextensive-explorationfeasible-reflectionknowledge-augmentationlarge-language-modelslonglong-chain-of-thoughtmultimodal-reasoningo1o3openai-o1overthinkingr1reasoningreasoning-language-modelsreinforcement-learningresearchrlsafetysurveysystem-2taxonomytest-time-scalingthinking