Evidence receipt / belief
Published · transcript-backedSholto Douglas: belief
28 Mar 2024 Dwarkesh Podcast Sholto Douglas & Trenton Bricken — How LLMs actually think
“I think directions like publishing the constitution that you expect your model to abide by–trying to make sure that you RLHF it towards that, and ablate that, and have the ability for everyone to offer feedback and contribution to that–is really important.”
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
…I think it depends on what government exactly has control and what the moral alignment is there. That is the whole value lock-in argument in my mind. It's definitely one of the strongest contributing factors for why I am working on capabilities at the moment. I think the current player set is actually extremely well-intentioned. For this kind of problem, I think we need to be extremely open about it. I think directions like publishing the constitution that you expect your model to abide by–trying to make sure that you RLHF it towards that, and ablate that, and have the ability for everyone to offer feedback and contribution to that–is really important. Sure. Alternatively, don't deploy when you're not sure. Which would also be bad because then we just never catch it.…
Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.