John Schulman - Reinforcement Learning from Human Feedback: Progress and Challenges

Name: John Schulman - Reinforcement Learning from Human Feedback: Progress and Challenges
Description: UC Berkeley EECS colloquium by OpenAI's John Schulman on reinforcement learning from human feedback. He covers why RL helps with truthfulness, hallucination, and the limits of behavior cloning.

UC Berkeley EECS

UC Berkeley EECS colloquium by OpenAI's John Schulman on reinforcement learning from human feedback. He covers why RL helps with truthfulness, hallucination, and the limits of behavior cloning.

Recommended in LLM threads as an authoritative explanation from a ChatGPT architect of why models hallucinate and why imitation is not enough.

5 Hacker News mentions · 2023–2023

Topics

Machine learning & deep learning
LLMs & generative AI

Watch on YouTube

Explore more videos