LLM Reasoning and Decomposition Techniques
13 repos
Methods for improving large language model reasoning through structured decomposition, chain-of-thought prompting, and compositional approaches. The cluster centers on techniques like breaking down complex problems into smaller steps, learning to reason through intermediate representations, and building better reasoning capabilities in transformer-based models. Repositories appear to implement and explore various prompting strategies and training approaches for enhancing LLM performance on reasoning tasks.
Jupyter Notebook · 1
Python · 1
