Institute for Explainable Machine Learning @HelmholtzMunich and Interpretable and Reliable Machine Learning group @TU_Muenchen
Munich, Germany
Joined August 2021
- Tweets400
- Following108
- Followers3.3K
- Likes210
Explainable Machine Learning retweeted
What does an LLM think of you?
LLMs pick up the slightest implicit cues about users. Often, they refuse to say who they think the user is.
In our new preprint, we show how to read and change these hidden beliefs. Turns out they even affect whether the model refuses
New paper🧵
Explainable Machine Learning retweeted
📣 New preprint: Replay on Demand (RoD)
Imagine you had 10 hours to refresh everything you learned in school.
Would you divide them equally across every subject?
Probably not. You’d spend your limited time on what you’ve forgotten most.
🧵 arxiv.org/abs/2609.40089
Explainable Machine Learning retweeted
📣 Introducing MedKIT, our new benchmark for knowledge integration in LLMs, accepted at #NeurIPS2026!
🧵 Why recalling updated facts ≠ integrated knowledge and why we need methods that can bridge this gap:
📄 arxiv.org/abs/2609.38543
Explainable Machine Learning retweeted
FineVision was just accepted to NeurIPS! Big thanks to all my co-authors, excited to present it at the conference
Explainable Machine Learning retweeted
Happy to have presented our work "FINER: MLLMs Hallucinate under Fine-grained Negative Queries" at #CVPR2026 as an oral (award candidate) and at #ICVSS2026
[Paper]: arxiv.org/abs/2603.17662
[Code]: github.com/ExplainableML/fin…
Explainable Machine Learning retweeted
Excited to share that our paper "SOTAlign: Semi-Supervised Alignment of Unimodal Vision and Language Models via Optimal Transport" was accepted at ICML.
📄 Paper: arxiv.org/abs/2602.23353
💻 Code: github.com/ExplainableML/SOT…
Explainable Machine Learning retweeted
A medical LLM can give the right answer for the wrong reasons.
Delighted our work is now published in Nature Communications Medicine — evaluating 8 LLMs on 1,269 unstructured seizure descriptions, with a practical guideline to use AI in clinical setup 🧵
doi.org/10.1038/s43856-026-0…
Explainable Machine Learning retweeted
Accurate reward guidance in diffusion models has been a long-standing challenge.
We propose Diamond Maps 💎, through stochastic flow maps we make reward guidance both accurate and scalable.
Explainable Machine Learning retweeted
🚨 New paper! arxiv.org/abs/2603.06610
Are you sure your post-trained LLM isn’t forgetting something?
Adapting LLMs is known to cause forgetting.
We usually measure it via general knowledge benchmarks.
But if MMLU doesn’t drop…. are you really fine?
Explainable Machine Learning retweeted
Check out our new work on getting more out of your Vision Transformers without fine-tuning by leveraging intermediate representations when probing!
We find that attentive probing is most robust for fusing across layers while showing exactly which layers matter for what task 👇
Explainable Machine Learning retweeted
Excited to be in San Diego at #NeurIPS2025 this week. We'll be presenting Noise Hypernetworks at Hall C,D,E #3605 Thu 4 Dec 11am!
If you're interested in Diffusion/Flow Matching or Reward Alignment, I'd love to chat! Feel free to DM me if you'd like to connect.
Explainable Machine Learning retweeted
Honored to receive one of the Google PhD Fellowships 2025 in ML Foundations for my research on "The Role of the Source Distribution in Generative Modeling: Beyond Fixed Gaussian Noise" :)
Grateful to @zeynepakata, Alexey, @TU_Muenchen, @Googleorg, and all my collaborators!
🎉 We're excited to announce the 2025 Google PhD Fellows! @GoogleOrg is providing over $10 million to support 255 PhD students across 35 countries, fostering the next generation of research talent to strengthen the global scientific landscape. Read more: goo.gle/43wJWw8
ALT A person with long hair is seen from behind working in a laboratory
Explainable Machine Learning retweeted
2 papers accepted at NeurIPS 2025 🎉
🔹 Manipulating Feature Visualizations with Gradient Slingshots
🔹 Capturing Polysemanticity with PRISM: A Multi-Concept Feature Description Framework
Explainable Machine Learning retweeted
Reward hacking is challenging when fine-tuning few-step Diffusion models. Direct fine-tuning on rewards can create artifacts that game metrics while degrading visual quality.
We propose Noise Hypernetworks as a theoretically grounded solution, inspired by test-time optimization.
Explainable Machine Learning retweeted
💫 After four PhD years on all things multimodal, pre- and post-training, I’m super excited for a new research chapter @GoogleDeepMind 🇨🇭!
Biggest thanks to @zeynepakata and @OriolVinyalsML for all the guidance, support, and incredibly eventful and defining research years ♥️!
Explainable Machine Learning retweeted
Personal news: I have defended my PhD thesis “Explaining Representations in Deep Neural Networks” at TU Berlin with summa cum laude (with distinction)!
From August, I’ll start a Postdoc at @TU_Muenchen in @ExplainableML group focusing on Mechanistic Interpretability ✨
Explainable Machine Learning retweeted
How can we circumvent data scarcity in the time series domain?
We propose to leverage pretrained ViTs (e.g., CLIP, DINOv2) for time series classification and outperform time series foundation models (TSFMs).
📄 Preprint: arxiv.org/abs/2506.08641
💻 Code: github.com/ExplainableML/TiV…
Explainable Machine Learning retweeted
I am very happy to announce that I successfully defended my PhD thesis with the title "Advancing Multimodal Explainability: From Visual Reasoning to In-Context Impersonation".
Explainable Machine Learning retweeted
Got a strong XAI paper rejected from ICCV? Submit it to our ICCV eXCV Workshop today—we welcome high-quality work!
🗓️ Submissions open until June 26 AoE.
📄 Got accepted to ICCV? Congrats! Consider our non-proceedings track.
#ICCV2025 @ICCVConference
Explainable Machine Learning retweeted
It is a great honor to receive the ZukunftsWissen Prize 2025 from the German Academy of the Sciences @Leopoldina with generous support of the @CoBaStiftung 🎉 This achievement wouldn’t have been possible without my wonderful group @ExplainableML @TU_Muenchen @HelmholtzMunich
🎉Der „ZukunftsWissen“-Preis 2025 von Leopoldina und @CoBaStiftung geht an @zeynepakata @TU_Muenchen/@HelmholtzMunich. Die Informatikerin erhält die mit 50.000 € dotierte Auszeichnung für ihre Forschung zur erklärbaren #KI.
leopoldina.org/presse-1/nach…
