Christian Internò (@ChrisInterno) header
Christian Internò (@ChrisInterno) profile picture

AI research @unibielefeld | Prev. @UniofOxford, @Honda Research, @CSHL, @LIACS

Joined September 2024
Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)Photo by Christian Internò (@ChrisInterno)
Christian Internò retweeted
Thrilled and honored to have been awarded a 2026 Google PhD Fellowship in Machine Learning and ML Foundations! My research proposal, “Foundations of Out-of-Distribution Discovery: from Theory to Molecules” aims to develop new theoretical and algorithmic principles toward a science of generative discovery. I'm deeply grateful to @GoogleResearch for this recognition and support, and to my PhD advisors @arkrause, @niao_he, and @kjelljorner, my Caltech mentors @yisongyue and @francesarnold, and my previous mentors Marcello Restelli, Marco Santambrogio, and @mmbronstein for their invaluable guidance and support. A huge thank you also to all the amazing collaborators, students, and colleagues I've had the privilege of working with and learning from along the way! Excited to keep working together!
15
5
96
4,154
Christian Internò retweeted
Thrilled to have been selected as a 2026 Google PhD Fellow in Machine Learning Foundations! Grateful to @GoogleResearch for supporting my research in AI × Causal Inference and its applications in experimental sciences.
We're excited to announce the 2026 recipients of the Google PhD Fellowship! 148 new PhD students across 38 countries, with $13M in support from @Googleorg. Their work spans quantum theory, language model reasoning, medical imaging and more: goo.gle/phd-fellows-2026
11
10
86
2,657
Christian Internò retweeted
Job update: AI Research Scientist at Aleph Alpha! Learning how LLMs are built in both industry and open source! (Kolibri - @Aleph__Alpha) (OpenEuroLLM - @OpenEuroLLM)
6
3
53
3,574
Christian Internò retweeted
Recent work on abstract synthetic data has shown that it can help pretrain LLMs more efficiently. But *what* does it actually improve? We identified gains in multi-hop/long-chain reasoning, and a series of conditions to obtain these benefits. 🧵
3
4
22
912
Christian Internò retweeted
Our first major update on icwm.cc - venue confirmed in the heart of Manhattan - paper deadline is December 22nd - single-blind, free-form (just a page limit), use AI however you want--focus on the science - everything public (openreview), < 3 papers per reviewer
10
35
8
351
36,985
Christian Internò retweeted
To me the real message of the bitter lesson was never that ideas don't matter. A good idea with little resources (compute, parameters, data) may lose to a bad idea with more resources. But the good ideas are the ones that give you the most bang for the buck. Good ideas improve scaling laws. They are the ones worth investing in. Good ideas are the ones that, resources being equal, beat bad ideas. So a good idea alone won't win you any prizes. It is when you differentiate ideas by resources that you see which ideas are good because they allow for better scaling. For example, rote memorization might lead to similar results as proper study on a single exam; but it will likely scale much worse over the course of an entire degree. This reminds me of @SchmidhuberAI 1991 noisy TV example. An agent that just tries to maximize prediction error will get stuck binging entropy. But an agent that maximizes the first temporal derivative of prediction error, i.e., learnability (glossing over some details) will be able to scale their knowledge over time. Don't give up on thinking hard and searching for great ideas. They will be the ones that scale to the moon 🌕
Brute force computation is not enough for intelligence. Clever new methods are needed. The science of AI is far from finished. IMHO
4
17
3
105
6,452
Christian Internò retweeted
I am not gonna comment on the results per se, but if you believe your work is valuable and insightful, why not tune the baselines? This took in total about 45 min of training on an A6000.
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: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
18
29
10
695
48,554
Christian Internò retweeted
I regret to inform you that regardless of how broken academia is, you are still responsible for what you become during your PhD. Your advisor can guide you. Your university can give you a degree. Neither can make you an independent thinker. Stop chasing paper counts. Stop waiting for the perfect research direction. Stop confusing publications with scientific progress. Stay curious. Question everything. Take ownership of your research. The goal was never to become a paper-producing machine. It was to become a scientist.
25
205
24
1,451
75,961
Christian Internò retweeted
In 2024, @zimmerrol and I wrote a paper on AutoInterp for vision models. The interpretability score (MIS) is basically an accuracy of how well most and least activating inputs can be distinguished. He analyzed over 70 Million neurons from over 800 models and found scores from 80% to over 90%. Back then, I was really into SAEs/Superposition and kind of bummed that even the neurons scored so high – maybe now this is relevant again? proceedings.neurips.cc/paper…
3
4
16
1,175
Christian Internò retweeted
Looking for a PhD program where you can study computational neuroscience? NYU has a fantastic array of researchers covering the field, and we'd love to consider your application: groups.google.com/g/systems-…
7
34
1
227
31,071
Christian Internò retweeted
Check out @hazel_heejeong latest paper. This was a pleasure to work on. Turns out now that we know World Models can be disentangled we can move on to Compositional Generalization: Has your model actually learned the dynamics of the world or is it just predicting a biased policy?
Your world model may know less about the world than you think. I've always wondered what world models really learn. Humans don't memorize every trajectory. We learn rules we can reuse. Can world models do the same? We built AnyGoal to find out. hazel-heejeong-nam.github.io…
3
40
2,169
Christian Internò retweeted
Your world model may know less about the world than you think. I've always wondered what world models really learn. Humans don't memorize every trajectory. We learn rules we can reuse. Can world models do the same? We built AnyGoal to find out. hazel-heejeong-nam.github.io…
13
27
4
186
11,101
Christian Internò retweeted
New surprising result: We trained AI recursively on its own safe generations, while giving it a benign cue. Models became increasingly misaligned over time, giving advice on how to do harm and build biological weapons. We call this Progressive Misalignment. 🧵👇
15
61
11
273
19,474
Christian Internò retweeted
Does this explain everything, like why we found direction selective neurons or face cells in the brain or why all these monosemantic neurons were/are about to be found in LLM MLPs? NO. This is theory. There are assumptions. It is always a bit of "theory fairyland" (@ChrisInterno). Our goal as theoreticians is to bring these assumptions as close as possible to reality (e.g., not just pretending the KL were the other way around). But ultimately, it is a model, hopefully a useful one that will spark further research, ideas and hypotheses ☺️
1
1
12
584
Christian Internò retweeted
Why? How did we ever get so lucky find meaningful neurons inside neural networks? We finally have a theory: arxiv.org/abs/2610.20851 🧵🧵🧵
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 already-clean.goodfire.com/
8
34
5
336
23,640
Christian Internò retweeted
In 2023 I showed @zimmerrol this neuron's most activating images and said that it is polysemantic (arxiv.org/abs/2310.11431). He said hold on, I think it codes for "crossing diagonal lines".. 🤯 That was eye opening and highlighted a limitation of our VisionAutoInterp metric (openreview.net/forum?id=oYyE…)
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 already-clean.goodfire.com/
3
10
2
145
7,031
Christian Internò retweeted
We’re bringing natural-language control to any scientific generator. We are very excited to introduce LLM-to-Flow Distillation (LFD), a method to adapt scientific generators to high-level natural-language discovery goals. Starting from textual goals like “design peptides with promising anticancer activity” or “design small molecules binding to the D2 dopamine receptor”, LFD distills high-level goal-related knowledge from LLMs into scientific generators through iterative DPO on LLM-generated sub-goals. 🔥 LFD does not need task-specific proxy rewards, works with any frontier LLM via API-access, and outperforms state-of-the-art reward-guided methods such as FlowGRPO and DiffusionNFT on held-out computational evaluations. (1/6)
7
41
7
175
26,942
Christian Internò retweeted
Looking for a PhD program where you can study vision from different perspectives (computer vision, visual neuroscience, visual perception, image processing)? Consider applying to NYU! visionscience.com/pipermail/…
1
19
1
161
8,686
Christian Internò retweeted
This is a result I’ve dreamt about for many years: *unpaired translation between images and text* I thought it might be only slightly possible, the kind of thing you have to really squint at. But Dominik proved this wrong. You don’t have to squint. Worth looking for yourself:
DINOv2 has never seen a caption, and Qwen3 has never seen an image. We still aligned their embedding spaces without a single image-caption pair. It even works when the images and the captions come from different datasets. Project page: dominik-schnaus.github.io/un… ⬇️
21
93
4
1,188
76,875
Christian Internò retweeted
What would a language model learn if its training data stopped before 1913? Our Language Model from 1913, TypewriterLM is headed to EMNLP 2026 🇭🇺!
3
8
487