1 paper · 1 filter
Md Nakhla Rafi, Md Ahasanuzzaman, Dong Jae Kim +2
LLM-based agents increasingly solve complex tasks through long trajectories involving reasoning steps, tool calls, and inter-agent communication. However, when these agents fail, i…