Alessandra Brondetta

PhD student | Inst. of Cognitive Science, Uni Osnabrück.

[Thomas Fel @thomas_fel_]

Happy to share my first post since joining Goodfire.

Neural geometry has been my obsession for years, and our team here is building a really serious research agenda around it.

I can't wait to share the series of papers coming over the next few weeks...

Brace for shapes 🍩

[Thomas Fel @thomas_fel_]

Thrilled to share that our work investigating recurrence in trained ViTs has been accepted at #ICLR2026! 🎉

It turns out DINOv2 may be secretly recurrent.🦖

A 2-block recurrent transformer can match its activations layer-by-layer and recover 96% of its ImageNet performance.

🧵

[Mozes Jacobs @mozesjacobs]

Are ViTs secretly RNNs? #ICLR2026

Our 2-block recurrent transformer recovers 96% of DINOv2’s IN-1k accuracy & reproduces its activations 1-to-1, motivating the Block-Recurrent Hypothesis: https://t.co/72E7FhY1LK

w/ @thomas_fel_ @RichieHakim @ABrondetta Demba Ba @t_andy_keller

[Thomas Fel @thomas_fel_]

🕳️🐇Into the Rabbit Hull – Part I (Part II tomorrow)

An interpretability deep dive into DINOv2, one of vision’s most important foundation models.

And today is Part I, buckle up, we're exploring some of its most charming features.

[Kaiqu Liang @kaiqu_liang]

🤔 Feel like your AI is bullshitting you? It’s not just you.

🚨 We quantified machine bullshit 💩

Turns out, aligning LLMs to be "helpful" via human feedback actually teaches them to bullshit—and Chain-of-Thought reasoning just makes it worse!

🔥 Time to rethink AI alignment.

[Ryan Liu @theryanliu]

A short 📹 explainer video on how LLMs can overthink in humanlike ways 😲!

had a blast presenting this at #icml2025 🥳

[Tom McCoy @RTomMcCoy]

🤖🧠 NEW PAPER ON COGSCI & AI 🧠🤖

Recent neural networks capture properties long thought to require symbols: compositionality, productivity, rapid learning

So what role should symbols play in theories of the mind? For our answer...read on!

Paper: https://t.co/VsCLpsiFuU

1/n

[Thomas Fel @thomas_fel_]

Train your vision SAE on Monday, then again on Tuesday, and you'll find only about 30% of the learned concepts match.

⚓ We propose Archetypal SAE which anchors concepts in the real data’s convex hull, delivering stable and consistent dictionaries.

https://t.co/iaX60GZt0o

[adithya @00aleph00]

The Goldbach conjecture isn’t as hard as you might think!

I know, I know - this sounds a bit scandalous. But I promise I’m not being disingenuous here.

There has actually been quite a lot of non-trivial progress towards it in the last 100 years.

(The Goldbach conjecture says that every even number bigger than 2 is the sum of two primes.)

A brief primer on the Goldbach conjecture and some things we know about it, a 🧵 1/n