Pierre Beckmann

Mechanistic interpretability and Philosophy of AI. PhD Student @epfl. Scholar @MATSProgram. Founder @philinterp.

[Pierre Beckmann @BeckmannPierre]

New paper: World Modeling in Transformers
with @matthieu_queloz @andrenfreitas @MATSprogram
Can AI models recover world models just from data? We find a faithful internal map and more inside of taxiGPT! We trace failures to how that the map is stored in superposition. 🧵(1/11)

[Pierre Beckmann @BeckmannPierre]

@ebervector @timfduffy Oh didn't see this before replying! Well I agree

[Pierre Beckmann @BeckmannPierre]

@timfduffy We found the same effect with the assistant axis (organises persona space)

https://t.co/moriE8kJs5

[Pierre Beckmann @BeckmannPierre]

Go read @FlaviohMar Escape dimensions theory if you might be interested in an upgrade to the lottery ticket hypothesis!

https://t.co/Xsv79mXxft

[Oliver @oliviscusAI]

🚨 MIT proved you can delete 90% of a neural network without losing accuracy.

The “lottery ticket hypothesis” found that some networks contain smaller subnetworks that can match the original model’s accuracy.

These “winning tickets” can be trained from their original starting weights. The finding does not apply to every model.

The challenge was finding them.

The original method required training the full network first, then pruning it and retraining the remaining connections.

That extra work made the approach harder to use in production.

Hardware support offers another way to benefit from sparsity.

NVIDIA Ampere and newer GPUs support a specific 2:4 pattern, where two weights out of every four are zero.

Compatible operations can use less memory bandwidth and reach up to twice the compute throughput.

But 2:4 sparsity removes 50% of those weights, not 90%. It also does not guarantee twice the speed for an entire model or zero accuracy loss.

Training methods and software support help turn fewer weights into practical savings. Results still depend on the model, hardware and pruning pattern.

The lesson is promising: some networks can achieve similar accuracy with far fewer connections.

Finding those connections and running them efficiently remain separate challenges.

[Jacob Charnock @charnock_jacob]

New GovAI post: I argue Frontier AI companies should publish their third-party evaluation contracts. This would show the conditions evaluators work under, make results easier to interpret and compare, and create incentives to improve evaluation arrangements.

More in thread 🧵

[Pierre Beckmann @BeckmannPierre]

@jayden_teoh_ What?? Very surprising and cool result

[Rikiya Takehi @rikiyatakehi]

We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.

We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: https://t.co/bOWCcFhCVq

[Benno Krojer @benno_krojer]

New preprint:
Cross-modal steering vectors transfer zero-shot in both directions 🖼️ ↔️ 📝
From text generation to image generation and vice-versa

Exciting evidence for a shared conceptual space!

Fantastic effort led by Donghao Zeng and awesome co-authors @nsubramani23 & @TamarRottShaham

Details 🧵👇

[Pierre Beckmann @BeckmannPierre]

fun stuff!

[max! @maxsloef]

You can turn Qwen3-8B into an owl by training only one random neuron! 🦉

Performance on owliness evals approaches full SFT. Compare SFT vs one-neuron eval responses side by side here: https://t.co/D8TiKeJbyx

[max! @maxsloef]

You can turn Qwen3-8B into an owl by training only one random neuron! 🦉

Performance on owliness evals approaches full SFT. Compare SFT vs one-neuron eval responses side by side here: https://t.co/D8TiKeJbyx

[Feijiang Han😈 @feijianghan]

check out this intuitive video on why interpretable directions in llms look non-gaussian, how ica finds them in minutes per layer, and the evaluation bug that made ica look weak for 3 years

[Pierre Beckmann @BeckmannPierre]

thanks for the shout-out!

[Mike Tamir, PhD @MikeTamir]

New research shows transformers do build internal world models, but localization failures stem from feature superposition. Mechanistic analysis reveals how representation and routing inte…

https://t.co/hZiKKr6Qgi

#MachineLearning #AI #LLM #DeepLearning #AgenticAI

[David Chalmers @davidchalmers42]

"three reasons why AI isn't conscious." really?

"consciousness is not intelligence": not a reason why AI isn't conscious, just rebuts a bad argument for AI consciousness (AI: intelligent therefore conscious).

"consciousness is substrate-dependent"/"noncomputational": would be a reason if true, but truth needs to be established.

more accurate title: "three reasons why AI might not be conscious".

https://t.co/rBmOZedWF3

[Sauers @Sauers_]

Language models can learn to detect when you read their activations. How? 🧵

[Emmanuel Ameisen @mlpowered]

What is an LLM?

Is it an incomprehensible black box, a next-word-predictor, a silicon genius?

To answer this, we don't have to only guess, we can open it up and look inside!

It was an honor to chat with @lennysan about what we learned doing this, some takeaways in 🧵

[Pierre Beckmann @BeckmannPierre]

@JeffYoshimi Yes!

[Pierre Beckmann @BeckmannPierre]

Here’s my take on the ongoing superposition debate that followed from this great release! I think it helps to distinguish structure-driven superposition from space-saving superposition. 1/

[Nick @nickcammarata]

interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had

with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works

https://t.co/zHij5zwKGT

[Pierre Beckmann @BeckmannPierre]

@trsam97 Here you go:
https://t.co/3DQDDoRxgK

[Pierre Beckmann @BeckmannPierre]

Here’s my take on the ongoing superposition debate that followed from this great release! I think it helps to distinguish structure-driven superposition from space-saving superposition. 1/

[Pierre Beckmann @BeckmannPierre]

Still, I don’t think this is a paradigm shift for mech interp. We’ll still need diff-means, probes and SAEs to find features (in particular for the res stream). But I’m excited to see more attribution graphs built directly on MLP neurons! 17/

[Pierre Beckmann @BeckmannPierre]

Overall, Nick gives evidence that there are fewer geometry tricks than we thought, at least for MLP neurons. More widely, this points toward superposition being driven more by structure than by geometry tricks. If true, latent spaces are indeed cleaner than prev. expected. 16/