The paper explores the potential of large multimodal models (LMMs) as generalist web agents that can complete tasks on websites. They propose a web agent called SEEACT and evaluate its performance on the MIND2WEB benchmark. The results show that LMMs like GPT-4V have the potential to complete tasks on live websites, but grounding remains a challenge.

https://arxiv.org/abs//2401.01614

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers