Go-Browse: Training Web Agents with Structured Exploration
Abstract
Go-Browse method collects web agent data through structured graph search, fine-tuning a 7B parameter language model to achieve high performance on the WebArena benchmark.
One of the fundamental problems in digital agents is their lack of understanding of their environment. For instance, a web browsing agent may get lost in unfamiliar websites, uncertain what pages must be visited to achieve its goals. To address this, we propose Go-Browse, a method for automatically collecting diverse and realistic web agent data at scale through structured exploration of web environments. Go-Browse achieves efficient exploration by framing data collection as a graph search, enabling reuse of information across exploration episodes. We instantiate our method on the WebArena benchmark, collecting a dataset of 10K successful task-solving trajectories and 40K interaction steps across 100 URLs. Fine-tuning a 7B parameter language model on this dataset achieves a success rate of 21.7% on the WebArena benchmark, beating GPT-4o mini by 2.4% and exceeding current state-of-the-art results for sub-10B parameter models by 2.9%.
Models citing this paper 2
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper