Adam Shai @adamimos

Joined June 2022

Tweets

180
Followers

447
Following

407
Likes

559

Ryan Kidd @ryan_kidd44

a week ago

MATS 9.0 applications are open! Launch your career in AI alignment, governance, and security with our 12-week research program. MATS provides field-leading research mentorship, funding, Berkeley & London offices, housing, and talks/workshops with AI experts.

13 60 264 1.0M 115

Download Image

Daniel Filan @dfrsrchtwts

4 days ago

🚨: if you apply to MATS, there is no guarantee that you won't be transformed into a 🌈lovely colourful geometric figure🌈. Behold the past fate of mentees of @danielmurfet, @jesse_hoogland, @RiechersPaul, and @adamimos. Apply to their streams, and the same fate could befall you!

2 8 103 6K 37

Download Image

Adam Shai @adamimos

3 days ago

Vibecoding on an airplane is about an order of magnitude more fun than my expectation. This is the life!!

0 0 5 215 0

Astera Institute @AsteraInstitute

a month ago

What’s going on inside large AI models? Astera grantees @adamimos and @RiechersPaul are building a new theory of internal structure to better understand intelligence. We sat down with them to learn more about their work as co-founders of Simplex, a research organization:…

1 4 23 8K 11

Download Image

Daniel Murfet @danielmurfet

a month ago

Neural networks are grown, not programmed. What does that growth process look like? Like this! This is a small language model (3M) across training, visualised with a new interpretability technique: susceptibilities. We call this handsome critter the rainbow serpent.

19 140 1K 80K 835

Download Image

Connor Leahy @NPCollapse

a month ago

@bayeslord Same bottlenecks as biology: Epistemology. It's barely a science, and getting it to the point where it is one is closer to philosophy or programming language design/UX than it is physics or math.

3 2 20 1K 0

Adam Shai @adamimos

a month ago

What a cool opportunity!

Oliver Habryka @ohabryka

a month ago

What a cool opportunity!

17 49 416 84K 173

Download Image

0 1 7 624 0

gavin leech @g_leech_

a month ago

The fractal activation geometry guys are back with even cooler (*superficially* dodgy-sounding!) results. No mention of this on twitter, surprisingly.

4 2 14 1K 3

Download Image

t-frex @tfrexbex

a month ago

And it’s pretty

Adam Shai @adamimos

a month ago

And it’s pretty

2 15 121 9K 82

Download Image

1 1 3 365 0

Jake Ward @_jake_ward

2 months ago

Do reasoning models like DeepSeek R1 learn their behavior from scratch? No! In our new paper, we extract steering vectors from a base model that induce backtracking in a distilled reasoning model, but surprisingly have no apparent effect on the base model itself! 🧵 (1/5)

4 27 253 34K 207

Download Video

Adam Shai @adamimos

2 months ago

What are the most beautiful research blogs presenting technical work? I'm a big fan of how Anthropic presents their transformer circuits work. Interested in others.