High Signal Podcasts Evidence ledger
Method
Browse
← Back to evidence

Evidence receipt / belief

Published · transcript-backed

Shawn Wang: belief

19 Jul 2023 Latent Space Llama 2: The New Open LLM SOTA (ft. Nathan Lambert, Matt Bornstein, Anton Troynikov, Russell Kaplan, Whole Mars Catalog et al.)

“I think that the Chinese universities have, have made some interesting progress there.”

— Shawn Wang

Source trail

Everything needed to verify it.

Speaker
Shawn Wang
Attribution
Verified speaker
Claim type
belief
Recorded
19 Jul 2023
Publisher
Latent Space

Transcript context

…One thing that I was thinking about is in the benchmarks they compare to G B T for, but if what George Hotz said on the podcast was right and should be D four is like eight attention heads. I wonder when people are gonna get eight, you know, get a LAMA two mixer expert going and benchmarking that. Maybe it will be better. I don't know. Yes, there, there is a little bit of a playbook that has been published out there, so I mean, it, it takes more skill than I, I have, but I'm sure someone else, else out there is currently working on it. I think that the Chinese universities have, have made some interesting progress there. Yeah, Simon, and then we'll go to Mars. So we talked about the we talked about retrieve augmented generation. The other thing I'm excited about is tool format, right? The the thing where it can call functions, essentially Uhhuh and that's mentioned in the paper. They mentioned they benchmarked along that, but, but I didn't get a feel for something that was really good at, the thing I want is I want basically exactly the same APIs, open AI functions, but I want it to run off of Llama2. I think that would be, that would open up all sorts of opportunities.…

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

Search evidence