AI Research Circle: AI Alignment, Unpacked

This session explores AI alignment through Anthropic's Constitutional AI framework. We'll dig into the 84-page constitution that guides Claude's behavior, the difference between rules and dispositions, and how Constitutional AI compares to RLHF. Co-facilitated by Anup Gosavi and Emily Hough-Kovacs. Pre-read: "Constitutional AI: Harmlessness from AI Feedback" (Bai et al., 2022).