1 paper · 1 filter
Zhichao Yang, Zhaoxin Fan, Gen Li +6
Structured, procedural reasoning is essential for Large Language Models (LLMs), especially in mathematics. While post-training methods have improved LLM performance, they still fal…