Lukas Muttenthaler
@lukasmut.bsky.social
Senior Researcher in AI for Biotech & @eml-munich.bsky.social | Prev. SR at Google DeepMind | PhD in ML and NeuroAI from @tuberlin.bsky.social @bifold.berlin @mpicbs.bsky.social | Representations in 🧠 and 🤖 | #FirstGen 🌈
💻: https://lukasmut.github.io/
💻: https://lukasmut.github.io/
Pinned
Lukas Muttenthaler
@lukasmut.bsky.social
· Nov 12
What aspects of human knowledge do vision models like CLIP fail to capture, and how can we improve them? We suggest models miss key global organization; aligning them makes them more robust. Check out LukasMuttenthaler's work, finally out (in Nature!?) www.nature.com/articles/s41... + our blog! 1/3
Aligning machine and human visual representations across abstraction levels - Nature
Aligning foundation models with human judgments enables them to more accurately approximate human behaviour and uncertainty across various levels of visual abstraction, while additionally improving th...
www.nature.com
🥳 I am incredibly humbled and grateful to share that our work, "Aligning machine and human visual representations across abstraction levels," has been published today in @nature.com ⬇️
🚨Go check out our most recent preprint on multilayer attention probing for Vision Transformers! It’s almost as performant as full fine-tuning while being similarly compute efficient as standard linear probing! Plus you get interpretable attention maps! More in 🧵👇🏼
Preprint: arxiv.org/abs/2601.09322
Preprint: arxiv.org/abs/2601.09322
January 19, 2026 at 9:46 AM
🚨Go check out our most recent preprint on multilayer attention probing for Vision Transformers! It’s almost as performant as full fine-tuning while being similarly compute efficient as standard linear probing! Plus you get interpretable attention maps! More in 🧵👇🏼
Preprint: arxiv.org/abs/2601.09322
Preprint: arxiv.org/abs/2601.09322
Even though I am not part of the Re-Align organizer team anymore, I am still cheering for this workshop! I still (and will always) believe that Re-Align is a great place for connecting the fields of AI / ML, Neuroscience, and Cognitive Science :) Can highly recommend submitting! 🧠🤖
🎉 Re-Align is back for its 4th edition at ICLR 2026!
📣 We invite submissions on representational alignment, spanning ML, Neuroscience, CogSci, and related fields.
📝 Tracks: Short (≤5p), Long (≤10p), Challenge (blog)
⏰ Deadline: Feb 5, 2026 for papers
🔗 representational-alignment.github.io/2026/
📣 We invite submissions on representational alignment, spanning ML, Neuroscience, CogSci, and related fields.
📝 Tracks: Short (≤5p), Long (≤10p), Challenge (blog)
⏰ Deadline: Feb 5, 2026 for papers
🔗 representational-alignment.github.io/2026/
January 7, 2026 at 9:03 PM
Even though I am not part of the Re-Align organizer team anymore, I am still cheering for this workshop! I still (and will always) believe that Re-Align is a great place for connecting the fields of AI / ML, Neuroscience, and Cognitive Science :) Can highly recommend submitting! 🧠🤖
Reposted by Lukas Muttenthaler
New post! Last week I shared why I thought cognitive (neuro)science hasn’t contributed as much as one might hope to the design of AI systems; this week I'm sharing my thoughts on how methods and principles from these fields *have* been useful in my work. infinitefaculty.substack.com/p/how-cognit...
How cognitive science can contribute to AI: methods for understanding
#2 in a series on cognitive science and AI
infinitefaculty.substack.com
December 23, 2025 at 5:10 PM
New post! Last week I shared why I thought cognitive (neuro)science hasn’t contributed as much as one might hope to the design of AI systems; this week I'm sharing my thoughts on how methods and principles from these fields *have* been useful in my work. infinitefaculty.substack.com/p/how-cognit...
So well deserved! 💛
December 7, 2025 at 7:56 PM
So well deserved! 💛
Reposted by Lukas Muttenthaler
What aspects of human knowledge do vision models like CLIP fail to capture, and how can we improve them? We suggest models miss key global organization; aligning them makes them more robust. Check out LukasMuttenthaler's work, finally out (in Nature!?) www.nature.com/articles/s41... + our blog! 1/3
Aligning machine and human visual representations across abstraction levels - Nature
Aligning foundation models with human judgments enables them to more accurately approximate human behaviour and uncertainty across various levels of visual abstraction, while additionally improving th...
www.nature.com
November 12, 2025 at 4:50 PM
What aspects of human knowledge do vision models like CLIP fail to capture, and how can we improve them? We suggest models miss key global organization; aligning them makes them more robust. Check out LukasMuttenthaler's work, finally out (in Nature!?) www.nature.com/articles/s41... + our blog! 1/3
Reposted by Lukas Muttenthaler
We summarize the ideas in a new DeepMind blog post: deepmind.google/blog/teachin... 2/3
Teaching AI to See the World More Like Humans Do
Aligning AI vision models with human knowledge, improves their robustness and ability to generalize.
deepmind.google
November 12, 2025 at 4:50 PM
We summarize the ideas in a new DeepMind blog post: deepmind.google/blog/teachin... 2/3
Reposted by Lukas Muttenthaler
When #AI “thinks” like us - New research out now in @nature.com from @lukasmut.bsky.social & @lampinen.bsky.social & @tuberlin.bsky.social & @mpib-berlin.bsky.social forms a bridge between #human and #machine representation. tinyurl.com/nhd4ycsh
When AI “thinks” like us
New research forms a bridge between human and machine representation
tinyurl.com
November 13, 2025 at 10:22 AM
When #AI “thinks” like us - New research out now in @nature.com from @lukasmut.bsky.social & @lampinen.bsky.social & @tuberlin.bsky.social & @mpib-berlin.bsky.social forms a bridge between #human and #machine representation. tinyurl.com/nhd4ycsh
🥳 I am incredibly humbled and grateful to share that our work, "Aligning machine and human visual representations across abstraction levels," has been published today in @nature.com ⬇️
What aspects of human knowledge do vision models like CLIP fail to capture, and how can we improve them? We suggest models miss key global organization; aligning them makes them more robust. Check out LukasMuttenthaler's work, finally out (in Nature!?) www.nature.com/articles/s41... + our blog! 1/3
Aligning machine and human visual representations across abstraction levels - Nature
Aligning foundation models with human judgments enables them to more accurately approximate human behaviour and uncertainty across various levels of visual abstraction, while additionally improving th...
www.nature.com
November 12, 2025 at 5:05 PM
🥳 I am incredibly humbled and grateful to share that our work, "Aligning machine and human visual representations across abstraction levels," has been published today in @nature.com ⬇️
Reposted by Lukas Muttenthaler
Pleased to share that our survey "Getting aligned on representational alignment" — on representational alignment across cognitive (neuro)science and machine learning — is now published in TMLR! openreview.net/forum?id=Hiq...
Kudos to @sucholutsky.bsky.social @lukasmut.bsky.social for leading this!
Kudos to @sucholutsky.bsky.social @lukasmut.bsky.social for leading this!
October 29, 2025 at 5:23 PM
Pleased to share that our survey "Getting aligned on representational alignment" — on representational alignment across cognitive (neuro)science and machine learning — is now published in TMLR! openreview.net/forum?id=Hiq...
Kudos to @sucholutsky.bsky.social @lukasmut.bsky.social for leading this!
Kudos to @sucholutsky.bsky.social @lukasmut.bsky.social for leading this!
Reposted by Lukas Muttenthaler
Though it’s traumatic for me to bring this up, I was targeted by a group allied with Turning Point USA in 2018, with the help of Tucker Carlson on Fox. My home address circulated online and I got death threats. 1/
September 13, 2025 at 3:21 PM
Though it’s traumatic for me to bring this up, I was targeted by a group allied with Turning Point USA in 2018, with the help of Tucker Carlson on Fox. My home address circulated online and I got death threats. 1/
Reposted by Lukas Muttenthaler
I’m recruiting committee members for the Technical Program Committee at #CCN2026.
Please apply if you want to help make submission, review & selection of contributed work (Extended Abstracts & Proceedings) more useful for everyone! 🌐
Helps to have: programming/communications/editorial experience.
Please apply if you want to help make submission, review & selection of contributed work (Extended Abstracts & Proceedings) more useful for everyone! 🌐
Helps to have: programming/communications/editorial experience.
Fill out the form here to be involved in #CCN2026
tinyurl.com/ccn26committee
And definitely fill out the #CCN2025 feedback form!
tinyurl.com/ccn25partici...
tinyurl.com/ccn26committee
And definitely fill out the #CCN2025 feedback form!
tinyurl.com/ccn25partici...
CCN 2025 Participant Survey
❗ This is the general CCN post-conference survey. For the application form for positions on the committees at CCN 2026, please see the CCN 2026 Committee Application.
Thank you for participating in C...
tinyurl.com
August 19, 2025 at 2:12 PM
I’m recruiting committee members for the Technical Program Committee at #CCN2026.
Please apply if you want to help make submission, review & selection of contributed work (Extended Abstracts & Proceedings) more useful for everyone! 🌐
Helps to have: programming/communications/editorial experience.
Please apply if you want to help make submission, review & selection of contributed work (Extended Abstracts & Proceedings) more useful for everyone! 🌐
Helps to have: programming/communications/editorial experience.
Reposted by Lukas Muttenthaler
Interesting reaction from ChatGPT to the HHS mRNA memo. It finds it so implausible that it thinks it's fake. From the perspective of a ~2024(?) trained model, 2025 policies are so absurd as to be unbelievable...
chatgpt.com/share/689364...
chatgpt.com/share/689364...
August 6, 2025 at 2:40 PM
Interesting reaction from ChatGPT to the HHS mRNA memo. It finds it so implausible that it thinks it's fake. From the perspective of a ~2024(?) trained model, 2025 policies are so absurd as to be unbelievable...
chatgpt.com/share/689364...
chatgpt.com/share/689364...
Reposted by Lukas Muttenthaler
🎉 Presenting at #ICML2025 tomorrow!
Come and explore how representational similarities behave across datasets :)
📅 Thu Jul 17, 11 AM-1:30 PM PDT
📍 East Exhibition Hall A-B #E-2510
Huge thanks to @lorenzlinhardt.bsky.social, Marco Morik, Jonas Dippel, Simon Kornblith, and @lukasmut.bsky.social!
Come and explore how representational similarities behave across datasets :)
📅 Thu Jul 17, 11 AM-1:30 PM PDT
📍 East Exhibition Hall A-B #E-2510
Huge thanks to @lorenzlinhardt.bsky.social, Marco Morik, Jonas Dippel, Simon Kornblith, and @lukasmut.bsky.social!
July 16, 2025 at 9:07 PM
🎉 Presenting at #ICML2025 tomorrow!
Come and explore how representational similarities behave across datasets :)
📅 Thu Jul 17, 11 AM-1:30 PM PDT
📍 East Exhibition Hall A-B #E-2510
Huge thanks to @lorenzlinhardt.bsky.social, Marco Morik, Jonas Dippel, Simon Kornblith, and @lukasmut.bsky.social!
Come and explore how representational similarities behave across datasets :)
📅 Thu Jul 17, 11 AM-1:30 PM PDT
📍 East Exhibition Hall A-B #E-2510
Huge thanks to @lorenzlinhardt.bsky.social, Marco Morik, Jonas Dippel, Simon Kornblith, and @lukasmut.bsky.social!
Reposted by Lukas Muttenthaler
If two models are more similar to each other than a third on ImageNet, will this hold for medical/satellite images?
Our #icml2025 paper analyses how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7
Our #icml2025 paper analyses how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7
June 6, 2025 at 2:14 PM
If two models are more similar to each other than a third on ImageNet, will this hold for medical/satellite images?
Our #icml2025 paper analyses how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7
Our #icml2025 paper analyses how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7
Reposted by Lukas Muttenthaler
What makes humans similar or different to AI? In a paper out in @natmachintell.nature.com led by @florianmahner.bsky.social & @lukasmut.bsky.social, w/ Umut Güclü, we took a deep look at the factors underlying their representational alignment, with surprising results.
www.nature.com/articles/s42...
www.nature.com/articles/s42...
Dimensions underlying the representational alignment of deep neural networks with humans - Nature Machine Intelligence
An interpretability framework that compares how humans and deep neural networks process images has been presented. Their findings reveal that, unlike humans, deep neural networks focus more on visual ...
www.nature.com
June 23, 2025 at 8:03 PM
What makes humans similar or different to AI? In a paper out in @natmachintell.nature.com led by @florianmahner.bsky.social & @lukasmut.bsky.social, w/ Umut Güclü, we took a deep look at the factors underlying their representational alignment, with surprising results.
www.nature.com/articles/s42...
www.nature.com/articles/s42...
Reposted by Lukas Muttenthaler
I am deeply grateful to @lorenzlinhardt.bsky.social, Marco Morik, Jonas Dippel, Simon Kornblith, and @lukasmut.bsky.social for their great work and support in this project! We also thank our collaborators, @bifold.berlin and HFA 7/7
📄Paper: arxiv.org/abs/2411.05561
💻Code: github.com/lciernik/sim...
📄Paper: arxiv.org/abs/2411.05561
💻Code: github.com/lciernik/sim...
Objective drives the consistency of representational similarity across datasets
The Platonic Representation Hypothesis claims that recent foundation models are converging to a shared representation space as a function of their downstream task performance, irrespective of the obje...
arxiv.org
June 6, 2025 at 2:14 PM
I am deeply grateful to @lorenzlinhardt.bsky.social, Marco Morik, Jonas Dippel, Simon Kornblith, and @lukasmut.bsky.social for their great work and support in this project! We also thank our collaborators, @bifold.berlin and HFA 7/7
📄Paper: arxiv.org/abs/2411.05561
💻Code: github.com/lciernik/sim...
📄Paper: arxiv.org/abs/2411.05561
💻Code: github.com/lciernik/sim...
🚨Have a look at our most recent work on the driving factors of representational similarities! This work will be presented at ICML 2025! See thread below 🧵
If two models are more similar to each other than a third on ImageNet, will this hold for medical/satellite images?
Our #icml2025 paper analyses how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7
Our #icml2025 paper analyses how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7
June 6, 2025 at 2:21 PM
🚨Have a look at our most recent work on the driving factors of representational similarities! This work will be presented at ICML 2025! See thread below 🧵
Please everyone gather and go outside, protest, march, do something! We cannot let science and academia be destroyed by the worst administration the US has ever had.
Breaking News: The Trump administration halted Harvard’s ability to enroll international students, a major escalation in its battle with the school.
Trump Administration Halts Harvard’s Ability to Enroll International Students
www.nytimes.com
May 22, 2025 at 8:06 PM
Please everyone gather and go outside, protest, march, do something! We cannot let science and academia be destroyed by the worst administration the US has ever had.
Reposted by Lukas Muttenthaler
How do language models generalize from information they learn in-context vs. via finetuning? In arxiv.org/abs/2505.00661 we show that in-context learning can generalize more flexibly, illustrating key differences in the inductive biases of these modes of learning — and ways to improve finetuning. 1/
arxiv.org
May 2, 2025 at 5:02 PM
How do language models generalize from information they learn in-context vs. via finetuning? In arxiv.org/abs/2505.00661 we show that in-context learning can generalize more flexibly, illustrating key differences in the inductive biases of these modes of learning — and ways to improve finetuning. 1/
Reposted by Lukas Muttenthaler
That’s a wrap for Re²-Align 🤖 🧠👤 at #ICLR2025 🦁🥳— thanks for being part of it! We’re probably just getting started… see you all at Re³-Align🌷😉
April 28, 2025 at 3:09 PM
That’s a wrap for Re²-Align 🤖 🧠👤 at #ICLR2025 🦁🥳— thanks for being part of it! We’re probably just getting started… see you all at Re³-Align🌷😉
Reposted by Lukas Muttenthaler
Harvard has set an example for other higher-ed institutions - rejecting an unlawful and ham-handed attempt to stifle academic freedom, while taking steps to make sure students can benefit from an environment of intellectual inquiry, rigorous debate and mutual respect. Let’s hope others follow suit.
April 15, 2025 at 3:52 AM
Harvard has set an example for other higher-ed institutions - rejecting an unlawful and ham-handed attempt to stifle academic freedom, while taking steps to make sure students can benefit from an environment of intellectual inquiry, rigorous debate and mutual respect. Let’s hope others follow suit.
The trump administration is the worst that could have happened.
April 17, 2025 at 4:47 PM
The trump administration is the worst that could have happened.