How to approach post-training for AI applications
January 17, 2025
6,638
213
4
3.27%
Search the Record
IndexedEvery word spoken in this episode is indexed. Type any phrase to jump straight to the moment it was said.
Type any word or phrase that may have been spoken. Click a result to seek the player to that exact moment.
Try a name, a topic, or a quoted line
Nathan Lambert Episodes Around January 17, 2025
See what was published immediately before and after this episode.
22:04Now PlayingHow to approach post-training for AI applications
YouTube Description
as posted by the channelMy talk during NeurIPs at Infer the Vancouver AI Engineering group
This was a fun one. I was trying to think of "what to say" to AI engineers.
What are the things I'm learning that actually translates to useful advice?
What will people building AI applications get wrong if they see new RLHF or finetuning papers and think they're going to try it?
When will we have a research ecosystem around finetuning from instruct?
This is a fun one, I hope you like it. As usual, reach out if you have questions!
(This also has a bunch of my content on OpenAI's reinforcement finetuning API)
For more, subscribe here and to my primary distribution channel, Interconnects.ai.
Get Interconnects ()...
... on YouTube
... on Twitter
... on Linkedin
... on Spotify
… on Apple Podcasts
Guests & Subjects Covered
Sentinel Indexing in Progress
Metadata and chapters are available. Claim extraction for this episode is pending.
All video content is delivered via YouTube embedded players in accordance with the YouTube Terms of Service. Sentinel provides research tools that promote discovery and accountability across political media.