RLF (Reinforcement Learning from Human Feedback) was the critical innovation that allowed ChatGPT to outperform Google's prior AI research; RLF involved humans rating responses (good paragraph vs. bad paragraph), and this human signal during training was the final breakthrough needed for the current generation of AI systems.
causalpending
Speaker
Eric SchmidtEvidence Quote
“there's a technology called rhf which involved humans adjusting the answers that was pioneered in jot GPT and that was the mining the the final thing that was needed the invention of rlf allowed the this the way it technically worked was it would give a human a choice of this paragraph or that paragraph and then human would say good paragraph bad paragraph and then the system would learn from that that additional human signal was key”
Created: 8/11/2026, 5:23:07 AM
Your notes
Loading notes...