working on AGI alignment. prev: GPT-Neo, the Pile, LM evals, RL overoptimization, scaling SAEs to GPT-4, interp via circuit sparsity. EleutherAI cofounder.
Leo Gao is an AI researcher working on AGI alignment, interpretability, and scaling sparse autoencoders at OpenAI. He co-founded EleutherAI and contributed to GPT-Neo, The Pile, and LM evals. Tagged AI Safety because alignment and safety research dominate his public identity.
Public AI-ranked followers appear first; other tracked followers remain below.
Last updated 15d ago · 1989 influencers tracked · 3000 ranked users