weakly typed @weakly_typed

learning {ML, PL, maths} // CS pre-grad // DMs open :) = fix 📍 Joined December 2021

Tweets

329
Followers

238
Following

555
Likes

3K

Neel Nanda @NeelNanda5

2 months ago

Exciting, mechanistic interpretability has a dedicated lecture in the syllabus of a Cambridge CS masters course! The field has come so far in the past few years ❤️

3 17 275 16K 67

Download Image

The slowly-unfolding premise of the Good Place is that everyone is damned. They are damned because they participate in the modern world; they buy from sweatshops, they eat chocolate, they fly in airplanes while the poorest people in the world see their harvests fail thanks to…

29 55 672 42K 223

Naomi Saphra @nsaphra

5 months ago

Take a break from arxiv/LW/AF. Sit in the woods with a random textbook and mull new ideas away from interp community lockstep. Diverge. Don’t compete with a saturated subtopic, maybe you’ll get to take weekends off. Premature overinvestment comes from monoculture.

Neel Nanda @NeelNanda5

5 months ago

1 3 69 36K 9

Download Image

5 18 220 29K 69

Zanzi Tangle, now at Monoidal Cafe @tangled_zans

8 months ago

I've recently learned about Algebraic Positional Encoding from @bgavran3 and isnt this the coolest breakthrough in mathematical approaches to transformers in the last few years arxiv.org/abs/2312.16045

4 20 179 9K 155

Mikel Bober-Irizar @mikb0b

9 months ago

LLMs are dramatically worse at ARC tasks the bigger they get. However, humans have no such issues - ARC task difficulty is independent of size. Most ARC tasks contain around 512-2048 pixels, and o3 is the first model capable of operating on these text grids reliably.

11 24 310 253K 75

Download Image

Samuel Marks @saprmarks

9 months ago

This is a really creative and well-executed paper on using "black-box interpretability" methods to understand and control model cognition. Especially impressed by the many applications explored IMO this is an important direction; this paper sets the field on an excellent path!

Alex Pan @aypan_17

9 months ago

4 26 138 21K 84

Download Image

1 2 23 3K 15

thebes @voooooogel

9 months ago

15 17 201 9K 24

Download Image

Jason Hausenloy @jasonhausenloy

10 months ago

The tragic suicide of Sewell Setzer III shows our generation has become unwitting test subjects in a vast, unregulated AI experiment. That's why we're launching @youthandai with our Generation AI Survey in @TIME. A thread: (1/10)

TIME @TIME

11 months ago

15 12 40 31K 2

2 11 24 4K 4

Transluce @TransluceAI

11 months ago

Announcing Transluce, a nonprofit research lab building open source, scalable technology for understanding AI systems and steering them in the public interest. Read a letter from the co-founders Jacob Steinhardt and Sarah Schwettmann: transluce.org/introducing-tr…

34 147 698 329K 256

Download Gif

weakly typed @weakly_typed

11 months ago

SHA-256: 218cebed21f2e8514df2ea1e4caca39750349cf30804995d5d577f08afc5855a

0 0 5 403 0

weakly typed @weakly_typed

a year ago

in slight defense of mathiness / mathematical notation in ML research papers: a thread (twessay?)

weakly typed @weakly_typed

a year ago

in slight defense of mathiness / mathematical notation in ML research papers: a thread (twessay?)

1 0 9 1K 0

0 1 5 1K 2

gavin leech @g_leech_

a year ago

Who should I meet in Cambridge? (You?)

4 3 17 3K 2

Allen Downey @AllenDowney

a year ago

On Reddit's statistics forum, the most common question is "What test should I use?" My answer, from 2011, is "There is only one test" allendowney.blogspot.com/2011/05/there-…

21 163 1K 204K 2K

Download Image

Jason Gross @diagram_chaser

a year ago

Mechanistic interpretability gives us rich explanations of models. But can we convert these explanations into formal proofs? Surprisingly, yes! Mech interp helps write short proofs of generalization bounds — and, shorter proofs provide more mechanistic understanding. 🧵

1 34 178 19K 130

Download Image

weakly typed @weakly_typed

a year ago

perhaps growing up is realising that 'growing up' was a comforting lie

2 0 3 339 0

weakly typed @weakly_typed

a year ago

on reading ml papers:

1 0 12 799 2

Download Image

weakly typed @weakly_typed

a year ago

maybe the most exciting interp result I’ve seen all year (if it ends up being true for interesting reasons): a meaningful step towards uncovering the type of the residual stream

Victor Veitch 🔸 @victorveitch

a year ago

maybe the most exciting interp result I’ve seen all year (if it ends up being true for interesting reasons): a meaningful step towards uncovering the type of the residual stream

11 121 619 80K 500

Download Video

0 0 4 307 0

henry @arithmoquine

a year ago

fyi the real reason i've been ignoring you is: - i want to reply - i want to be able to give you the attention and focus you deserve - i never feel like i have enough energy to properly do that

henry @arithmoquine

a year ago

fyi the real reason i've been ignoring you is: - i want to reply - i want to be able to give you the attention and focus you deserve - i never feel like i have enough energy to properly do that

4 1 44 6K 3

Download Image

2 3 93 8K 21

weakly typed @weakly_typed

a year ago

mechinterp people: does anyone have a good (formal?) definition of 'feature' that doesn't assume the linear representation hypothesis? like, if I have some points in high-dim space, what makes them "the composition of several features" as opposed to "some random points"