Skip to content

Instantly share code, notes, and snippets.

View tuan3w's full-sized avatar
👋

Tuan Nguyen tuan3w

👋
View GitHub Profile

Reinforcement Learning for Language Models

Yoav Goldberg, April 2023.

Why RL?

With the release of the ChatGPT model and followup large language models (LLMs), there was a lot of discussion of the importance of "RLHF training", that is, "reinforcement learning from human feedback". I was puzzled for a while as to why RL (Reinforcement Learning) is better than learning from demonstrations (a.k.a supervised learning) for training language models. Shouldn't learning from demonstrations (or, in language model terminology "instruction fine tuning", learning to immitate human written answers) be sufficient? I came up with a theoretical argument that was somewhat convincing. But I came to realize there is an additional argumment which not only supports the case of RL training, but also requires it, in particular for models like ChatGPT. This additional argument is spelled out in (the first half of) a talk by John Schulman from OpenAI. This post pretty much

@nickloewen
nickloewen / bret_victor-reading_list.md
Last active July 12, 2024 17:54
Bret Victor’s Reading List

This is a plain-text version of Bret Victor’s reading list. It was requested by hf on Hacker News.


Highly recommended things!

This is my five-star list. These are my favorite things in all the world.

A few of these works have had an extraordinary effect on my life or way of thinking. They get a sixth star. ★