News/Will the Agent Recuse Itself? Measuring LLM-Agen…
Satelight

Will the Agent Recuse Itself? Measuring LLM-Agent Compliance with In-Band Access-Deny Signals

Read full release on Satelight
  • OpenAI
  • Anthropic
Will the Agent Recuse Itself? Measuring LLM-Agent Compliance with In-Band Access-Deny Signals