High Signal Podcasts Evidence ledger
Method
Browse
← Back to evidence

Evidence receipt / belief

Published · transcript-backed

Sholto Douglas: belief

28 Mar 2024 Dwarkesh Podcast Sholto Douglas & Trenton Bricken — How LLMs actually think

“There’s one mode of thinking in which fine-tuning is specialized, you've got this latent bundle of capabilities and you're specializing it for this particular use case that you want. I think I'm not sure how true or not that is.”

— Sholto Douglas

Source trail

Everything needed to verify it.

Speaker
Sholto Douglas
Attribution
Verified speaker
Claim type
belief
Recorded
28 Mar 2024
Publisher
Dwarkesh Podcast

Transcript context

…Okay, interesting. I find the fact that fine-tuning works as evidence of curriculum learning, right? Because the last things you're training on have a disproportionate impact. I wouldn't necessarily say that. There’s one mode of thinking in which fine-tuning is specialized, you've got this latent bundle of capabilities and you're specializing it for this particular use case that you want. I think I'm not sure how true or not that is. I think the David Bell lab paper kind of supports this. You have that ability and you're just getting better at entity recognition, fine-tuning that circuit instead of other ones.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.

Search evidence