Web-Shepherd: Advancing PRMs for Reinforcing Web Agents
This work addresses the need for specialized, cost-effective reward models for web agents, which is important for automating repetitive real-life tasks, though it appears incremental as it builds on existing PRM concepts applied to a new domain.
The authors tackled the problem of web navigation automation by developing Web-Shepherd, the first process reward model (PRM) for step-level assessment of web navigation trajectories, which achieved about 30 points better accuracy than GPT-4o on their benchmark and 10.9 points better performance with 10× less cost in testing.
Web navigation is a unique domain that can automate many repetitive real-life tasks and is challenging as it requires long-horizon sequential decision making beyond typical multimodal large language model (MLLM) tasks. Yet, specialized reward models for web navigation that can be utilized during both training and test-time have been absent until now. Despite the importance of speed and cost-effectiveness, prior works have utilized MLLMs as reward models, which poses significant constraints for real-world deployment. To address this, in this work, we propose the first process reward model (PRM) called Web-Shepherd which could assess web navigation trajectories in a step-level. To achieve this, we first construct the WebPRM Collection, a large-scale dataset with 40K step-level preference pairs and annotated checklists spanning diverse domains and difficulty levels. Next, we also introduce the WebRewardBench, the first meta-evaluation benchmark for evaluating PRMs. In our experiments, we observe that our Web-Shepherd achieves about 30 points better accuracy compared to using GPT-4o on WebRewardBench. Furthermore, when testing on WebArena-lite by using GPT-4o-mini as the policy and Web-Shepherd as the verifier, we achieve 10.9 points better performance, in 10 less cost compared to using GPT-4o-mini as the verifier. Our model, dataset, and code are publicly available at LINK.