DialogGuard: Multi-Agent Psychosocial Safety Evaluation of Sensitive LLM Responses

Open in new window