← 返回飞轮报告库

LLM智能体行为异常的实时检测:基于‘语义一致性’和‘逻辑连贯性’的轻量级模型

Run ID: run-6de73533aceb Score: 0.81 Verdict: converged Generated: 2026-05-26