1
Behind the Refusal: Determining Guardrail Activation via Behavioral Monitoring
揭秘大模型“拒绝”背后的护栏激活机制：如何通过行为监控精准判断模型何时触发安全防线？
arXiv:2607.02121v1 Announce Type: cross Abstract: As Large Language Models (LLMs) and agentic systems become integrated into real-world applications, …