Reasoning using Language Models
FreeCollection of papers and resources on Reasoning using Language Models.
FreeFree tier
About Reasoning using Language Models
A curated collection of papers and resources on how to unlock the reasoning ability of Large Language Models (LLMs) and Multimodal LLMs (MLLMs). This repository covers chain-of-thought prompting, multimodal reasoning, scaling smaller models, and includes survey and analysis papers. It also provides a symbolic reasoning benchmark (LLMSymbolicReasoningBench) and is continually updated with contributions from the community.
Key Features
Curated collection of reasoning-focused LLM papers and resources
Covers chain-of-thought prompting, multimodal reasoning, and scaling smaller models
Organized into surveys, analysis, and technique sections
Includes links to code and additional resources (e.g., benchmarks)
Open to community contributions via pull requests
Pros & Cons
Pros
- Comprehensive and regularly updated collection
- Includes both survey and analysis papers from top conferences and preprints
- Free and open source with community contributions
- Provides links to code for many papers
Cons
- Not an interactive tool; requires manual browsing of the list
- May not include the very latest preprints until added
- No built-in search or filtering for nuanced topics
Best For
Academic research on LLM reasoning capabilitiesLiterature review for papers on chain-of-thought, o1, DeepSeek-R1, and related topicsUnderstanding and comparing reasoning techniques and benchmarksExploring symbolic reasoning benchmarks (LLMSymbolicReasoningBench)
FAQ
What does this repository cover?
It is a curated collection of papers and resources on reasoning in Large Language Models (LLMs) and Multimodal LLMs, covering surveys, analysis, techniques such as chain-of-thought prompting, and scaling smaller models.
How is the repository organized?
It is organized into tables of contents including Survey, Analysis, Technique, Reasoning in LLMs, Multimodal Reasoning, Scaling Smaller Language Models, and Other Useful Resources.
Can I contribute to this collection?
Yes, the repository is open to contributions. It includes a 'Contributing' section and accepts pull requests.