CHATS Lab 
Conversation, Human-AI Technology, Safety (CHATS) Lab studies how to
build chatbots and break chatbots with interdisciplinary approaches.
Example projects
- Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity
- Persuasive Jailbreaker: How to Persuade LLMs to Jailbreak It?
- KokoMind: Do LLM Understand Social Interactions?
Funding
We thanks the generous support of the following sponsors and organizations: Schmidt Sciences, Open Philanthropy, and Massachusetts AI Hub.
