[arXiv]score: 0.86
Agentic CLEAR: Automating Multi-Level Evaluation of LLM Agents
May 21, 2026
Agentic CLEAR is an automatic evaluation framework for LLM agents that produces dynamic, domain-adaptive textual insights at system, trace, and node levels without hand-crafted error taxonomies.
cs.CLcs.AI