RLHF (reinforcement learning from human feedback) unhobbled GPT-2's raw capabilities into a functional chatbot, achieving roughly a 100x improvement in human preference ratings compared to non-RLHF models, and enabled basic chain-of-thought reasoning, but the underlying model still possessed many unutilized raw capabilities that could be further unlocked.
factualpending
Speaker
Leopold AschenbrennerEvidence Quote
“The unhobbling from GPT-2 to GPT-4 took this raw mass and RLHF'd it into a good chatbot... comparing RLHF vs. non-RLHF models it's like a 100x model size win on human preference rating.”
Source
Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history— Dwarkesh PatelCreated: 8/12/2026, 6:39:27 PM
Your notes
Loading notes...