1 paper · 1 filter
Yufeng Peng, Yixin Zhang, Ziyi Liu +5
Large Language Model agents operate through closed-loop action policies whose decision space is defined by an agent harness. We investigate how this operating space shapes agent be…