SAGE-RT: Synthetic Alignment data Generation for Safety Evaluation and Red Teaming
abstract
We introduce SAGE-RT, a comprehensive framework for generating synthetic alignment data tailored for safety evaluation and red teaming of generative AI systems. Our approach addresses the critical need for diverse, high-quality evaluation datasets in AI safety research by providing automated generation of challenging test cases.
Red TeamingAI SafetySynthetic DataAlignmentEvaluation
cite
@inproceedings{kumar2024sage-rt,
title = {SAGE-RT: Synthetic Alignment data Generation for Safety Evaluation and Red Teaming},
author = {Anurakt Kumar and Divyanshu Kumar and Jatan Loya and Nitin Aravind Birur and Tanay Baswa and Sahil Agarwal and Prashanth Harshangi},
booktitle = {NeurIPS Red Teaming GenAI Workshop},
year = {2024}
}