LLM agent performance on anomaly detection and root-cause analysis tasks has not been systematically evaluated under controlled conditions

their performance on these tasks has not been systematically evaluated under controlled conditions

http://arxiv.org/abs/2608.27182v1