1 paper · 1 filter
Terumi Chiba, Guangzhi Sun, Zheqi Yuan +1
Large language model (LLM) computer-use agents are typically evaluated with clean written instructions, despite speech being an increasingly popular interface for interacting with…