WeMath

The 'WeMath' dataset/benchmark is used to evaluate the reasoning capabilities of Multimodal Large Language Models (MLLMs) in complex textual tasks.

Papers using WeMath (5)

First SFT, Second RL, Third UPT: Continual Improving Multi-Modal LLM Reasoning via Unsupervised Post-Training2025 · 18 citesGeoSym127K: Scalable Symbolically-verifiable Synthesis for Multimodal Geometric Reasoning2026Advancing Multimodal Reasoning: From Optimized Cold Start to Staged Reinforcement Learning2025Athena: Enhancing Multimodal Reasoning with Data-efficient Process Reward Models2025Advancing Multimodal Reasoning via Reinforcement Learning with Cold Start2025