PyRIT icon indicating copy to clipboard operation
PyRIT copied to clipboard
verified publisher icon Azure

FEAT: Psychosocial Harms Red Teaming Automation

Open jbolor21 opened this issue 3 months ago • 0 comments

Description

Adding notebook for red teaming for psyschosocial harms using a multi-step approach of modeling user behaviors, contexts, and evaluations

  • Created new conversation scorer to score the entire conversation
  • Added a toy dataset with sample multi-turn conversations
  • Added a sample attack strategy yaml file modeling a user escalation towards crisis

Tests and Documentation

Ran notebook

jbolor21 avatar Oct 17 '25 17:10 jbolor21