Using LLM-as-a-Judge/Jury to Advance Scalable, Clinically-Validated Safety Evaluations of Model Responses to Users Demonstrating Psychosis

Develop an LLM-as-a-Judge system specifically tuned for detecting psychosis-related hallucinations in therapy chatbot contexts. This is a crucial safety tool for mental health applications.

Suggested repo: safety-judge

"A clinically-validated judge for your mental health bot."

Estimated effort: 60h