WebArbiter
Collection
WebArbiter: Reasoning Process Reward Model for Web Agents. Models, training data, and WebPRMBench. ICLR 2026. • 4 items • Updated
This model will be released shortly. Stay tuned!
Paper: WebArbiter: A Principle-Guided Reasoning Process Reward Model for Web Agents
Code: GitHub
Website: yaozhang.ai/WebArbiter