Math, Code, and Reasoning in LLMs
36 repos
Systems and techniques for enhancing large language models' capabilities in mathematical reasoning, code generation, and logical problem-solving. This cluster covers instruction-tuning datasets (like MathCodeInstruct), retrieval-augmented approaches, and specialized model architectures designed to improve transformer-based text generation for technical domains. Projects here address the challenge of making LLMs more reliable and accurate when handling symbolic reasoning, mathematical proofs, and code synthesis tasks.
Python · 1
math ·3,847
reasoning ·2,946
code ·1,900
transformers ·1,017
en ·1,017
text-generation ·1,017
llm ·905
synthetic ·860
safetensors ·821
instruction ·798
