Skip to content
Trending storyDeveloping

Lewis Tunstall endorses SFT before expensive reinforcement learning

1 article1 sourcesince Nov 5Last article 1h ago ·

Overview

AISummary of 1 article

Lewis Tunstall of Hugging Face says he approves of the message that supervised fine-tuning, or SFT, should be used for everything.

He endorsed a post by Esong Ghori advising teams to get as far as possible with SFT before relying on expensive reinforcement learning, or RL.

Tunstall calls RL "the cherry on top" and cites Yann LeCun for the point. The report is a single X post, and it gives no performance figures or cost comparisons.

Written by AI from the articles below · updated Nov 5, 5:54 PM ET

Check the sources:

Latest developmentNov 5, 5:40 PM ET
Lewis Tunstall backs supervised fine-tuning over expensive reinforcement learning

Article timeline

The articles in this story. Times are ET.

Nov 5
  1. Lewis TunstallX
    Lewis Tunstall backs supervised fine-tuning over expensive reinforcement learning

    AILewis Tunstall of Hugging Face says he approves of the message that supervised fine-tuning, or SFT, should be used for everything. A quoted post from Esong Ghori advises climbing as far as possible with SFT before relying on expensive reinforcement learning, or RL. Tunstall adds that RL is the cherry on top, citing Yann LeCun.

Heat Trend

Not enough continuous observations to show a trend yet.