We-Math

Dataset Card for WE-MATH (ACL 2025) GitHub | Paper | Website Inspired by human-like mathematical reasoning, we introduce We-Math, the first benchmark specifically designed to explore the problem-solving principles beyond the end-to-end performance. We meticulously collect and categorize 6.5K visual math problems, spann

Papers using We-Math (3)

Advancing Multimodal Reasoning via Reinforcement Learning with Cold Start2025Unsupervised Post-Training for Multi-Modal LLM Reasoning via GRPO2025Advancing Multimodal Reasoning via Reinforcement Learning with Cold Start2025