Machine learningLLMs & Text

ChatGLM-Math: Improving Math Problem-Solving in Large Language Models with a Self-Critique Pipeline

Problem-Solving in LLMs: The authors propose a Self-Critique pipeline to enhance the mathematical problem-solving skills of large language models without affecting their language abilities, showing significant improvements in both areas.