1 paper · 1 filter
Haodong Lei, Junming Liu, Yirong Chen +4
Large language model (LLM) agents increasingly operate over long interaction histories, where effective reasoning requires identifying and exploiting task-relevant evidence distrib…