LLM Reasoning & Decomposition Techniques
2 repos
Research and implementation of advanced prompting and inference strategies for large language models, focusing on methods that break down complex reasoning tasks into manageable components. The cluster centers on techniques like chain-of-thought reasoning, step-by-step decomposition, and structured problem-solving approaches that improve LLM performance on multi-step tasks. Repositories here explore various frameworks for orchestrating model reasoning and evaluation of decomposed reasoning paths.
