Explaining AI Agents Through Execution Traces
This work presents a post-hoc XAI framework that transforms a lengthy agent's execution trace into a structured report and a faithful natural-language explanation explicitly grounded in its observable behavior, outperforming naive LLM-generated explanations.