iiinova — where your digital self keeps evolving

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 27/09/2026

This is what my research with colleagues at Aleph Alpha Research on Merlin-Arthur protocols is about. Here's a video to explain the complicated aspects of this work.

000

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 27/09/2026

New AI Coffee Break video! ☕️🔥 LLMs can hallucinate and give the correct answer for the wrong reason. What about teaching them to admit “I don’t know”? 📺 youtu.be/796GVFTFiB0

youtu.be

Reducing Hallucinations: Merlin-Arthur Training and Evaluation EXPLAINED

YouTube video by AI Coffee Break with Letitia

120

Reposted by aicoffeebreak.bsky.social

Heidelberg Laureate Forum @hlforum.bsky.social · 22/07/2026

💡Today, we speak with HLF alumna Letiția Pârcălăbescu @aicoffeebreak.bsky.social who will be joining us at the 13th HLF as a panelist! 🖥️🎓 During her PhD in computer science, Letitia developed tools to help multimodal LLMs complete tasks more "honestly". 👉 scilogs.spektrum.de/hlf/?p=14524

scilogs.spektrum.de

Keeping AI Honest - Heidelberg Laureate Forum - SciLogs - Wissenschaftsblogs

We speak to computer scientist Letiția Pârcălăbescu on her research to help train more transparent and "honest" AI models.

022

Reposted by aicoffeebreak.bsky.social

Women in AI Research - WiAIR @wiair.bsky.social · 18/02/2026

🧠 Do Vision & Language Decoders Use Images and Text Equally? In our latest episode, we speak with Letitia Parcalabescu about her ICLR 2025 paper examining how vision–language *decoder* models use images and text — and how self-consistent their explanations really are. (1/8🧵)

131

Reposted by aicoffeebreak.bsky.social

Women in AI Research - WiAIR @wiair.bsky.social · 06/02/2026

If you love @aicoffeebreak.bsky.social, this one's for you — Letitia Parcalabescu is our next guest on the #WiAIR_podcast! Stay tuned for our conversation: 🎬 YouTube: www.youtube.com/@WomeninAIRe...

121

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 02/11/2025

LLMs can memorize even a phone number seen once in training.🔒 Google’s VaultGemma fixes that, being the first open-weight LLM trained from scratch with differential privacy, so rare secrets leave no trace. ☕ new video explaining Differential Privacy through VaultGemma 👇 🎥 youtu.be/UwX5zzjwb_g

youtu.be

What's up with Google's new VaultGemma model? – Differential Privacy explained

YouTube video by AI Coffee Break with Letitia

062

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 19/10/2025

We explain diffusion models and flow-matching models side by side. Flow-Matching models are the new generation of AI image generators that are quickly replacing diffusion models. They take everything diffusion did well, but make it faster, smoother, and deterministic. 🎥 youtu.be/firXjwZ_6KI

youtu.be

Diffusion Models and Flow-Matching explained side by side

YouTube video by AI Coffee Break with Letitia

021

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 21/09/2025

Ever wondered how Energy-Based Models (EBMs) work and how they differ from normal neural networks? ☕️ We go over EBMs and then dive into the Energy-Based Transformers paper to make LLMs that refine guesses, self-verify, and could adapt compute to problem difficulty.

3226

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 14/09/2025

The world’s largest NLP conference with almost 2,000 papers presented, ACL 2025 just took place in Vienna! 🎓✨ Here is a quick snapshot of the event via a short interview with one of the authors whose work caught my attention. 🎥 Watch: youtu.be/GBISWggsQOA

032

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 05/08/2025

Check it out if you’re curious or feel like supporting thoughtful science storytelling; there's a “text trailer” on these pages 📜✨: 🔗 scifilmit.com/puppetsofadi...

scifilmit.com

Puppets of a Digital Brain – SciFilmIt

000

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 05/08/2025

My friend Vivi Nastase is working on a short science communication film called "Puppets of a Digital Brain". It aims to explain the tech behind AI chatbots (the good, the bad, the environmental) in an accessible, visual way. 💡 GoFundMe: gofund.me/453ed662

231

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 03/08/2025

In this video, we break down each method and show how the same model can sound dull, brilliant, or unhinged – just by changing how it samples. 🎥 Watch here: youtu.be/o-_SZ_itxeA

youtu.be

Greedy? Random? Top-p? How LLMs Actually Pick Words – Decoding Strategies Explained

YouTube video by AI Coffee Break with Letitia

000

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 03/08/2025

How do LLMs pick the next word? They don’t choose words directly: they only output word probabilities. 📊 Greedy decoding, top-k, top-p, min-p are methods that turn these probabilities into actual text.

132

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 27/07/2025

I'll co-organise the "Multilingualism: from data crawling to evaluation" social / birds-of-a-feather session. It's on the 29th at 4PM. Do come by if you're at ACL Vienna! :)

010

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 27/07/2025

I'm also at ACL, would be lovely to catch up!

120

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 27/07/2025

Right now, attending the synthetic gata generation tutorial which is packed, because it turned out that "Data is the new source code." #ACL2025NLP

020

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 27/07/2025

Excited to be at ACL 2025 in Vienna this week 🇦🇹 #ACL2025 I’m always up for a chat about reasoning models, NLE faithfulness, synthetic data generation, or the joys and challenges of explaining AI on YouTube. If you're around, let’s connect!

281

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 07/07/2025

📅 Looking forward to the discussion and to learning from fellow panelists and participants. If you're around Heidelberg, join us! www.marsilius-kolleg.uni-heidelberg.de/de/dancing-w...

marsilius-kolleg.uni-heidelberg.de

Dancing with Right & Wrong? - Marsilius-Kolleg

An interdisciplinary symposium on knowing, doing and not being sure

011

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 07/07/2025

🔨 What recent advancements in AI were particularly impactful for science? 🔍 How do we calibrate trust in current AI systems? 🧪 If AI takes over more of the scientific process… what’s left for us humans?

120

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 07/07/2025

🤖 Can we trust AI in science? I'm excited to be speaking at the final event of the Young Marsilius Fellows 2025, themed "Dancing with Right & Wrong?" – a title that feels increasingly relevant these days. I'll be joining a panel on "(How) can we trust AI in science?" to discuss questions like:

121

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 20/06/2025

We train AI on human-selected or -generated data (yes, even taking a photo is concept selection – we capture what we find interesting; text even more so, expressing our conceptualisation of the world). Then we’re surprised when the AI's concepts and representations are similar to ours. 🤷‍♀️

110

Reposted by aicoffeebreak.bsky.social

Nils Trost @trostnils.bsky.social · 20/06/2025

I'm very excited to finally share the main work of my PhD! We explored the evolutionary dynamics of gene regulation and expression during gonad development in primates. We cover among others: X chromosome dynamics (incl. in a developing XXY testis), gene regulatory networks and cell type evolution.

1145

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 19/06/2025

💡 AlphaEvolve is a new AI system that doesn’t just write code, it evolves it. It uses LLMs and evolutionary search to make scientific discoveries. We explain how AlphaEvolve works and the evolutionary strategies behind it (like MAP-Elites and island-based population methods). 📺 youtu.be/Z4uF6cVly8o

youtu.be

AlphaEvolve: Using LLMs to solve Scientific and Engineering Challenges | AlphaEvolve explained

YouTube video by AI Coffee Break with Letitia

020

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 31/05/2025

💡 Participation includes talks, workshops, and lots of cross-disciplinary exchange—with accommodation and meals covered (fee: 100€). If this sounds like your thing, the application deadline is June 27!

100

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 31/05/2025

📍 Heidelberg, September 21–27, 2025 💬 Language: English 🎯 Open to PhD students & advanced Master’s students from all disciplines working on AI-related research

110

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 31/05/2025

👉 www.marsilius-kolleg.uni-heidelberg.de/de/studium/i... 🧠 This interdisciplinary event brings together researchers from across fields—computer science, linguistics, philosophy, law, medicine, theology—to explore the normative foundations of generative AI, and how values are embedded in its design.

marsilius-kolleg.uni-heidelberg.de

AI and Human Values - Marsilius-Kolleg

100

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 31/05/2025

Excited to share that I’ll be joining the Summer School “AI and Human Values” this September at the Marsilius-Kolleg of Heidelberg University as a speaker. I'll be giving an introduction to how large language models actually work—before the summer school dives deeper into broader implications.

132

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 18/05/2025

Long videos are a nightmare for language models—too many tokens, slow inference. ☠️ We explain STORM ⛈️, a new architecture that improves long video LLMs using Mamba layers and token compression. Reaches better accuracy than GPT-4o on benchmarks and up to 8× more efficiency. 📺 youtu.be/uMk3VN4S8TQ

youtu.be

Token-Efficient Long Video Understanding for Multimodal LLMs | Paper explained

YouTube video by AI Coffee Break with Letitia

052

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 17/05/2025

Thank you! I post here as frequently as on Twitter. 🙈 I'm doing videos once a month now.

110

Reposted by aicoffeebreak.bsky.social

Evangelos Kazakos @ekazakos.bsky.social · 17/05/2025

Follow @aicoffeebreak.bsky.social!! Letitia is very effective in communicating research papers in just a few mins! Perfect for your coffee break. 😉

022

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 18/04/2025

We all know quantization works at inference time, but researchers successfully trained a 13B LLaMA 2 model using FP4 precision (only 16 values per weight!). 🤯 We break down how it works. If quantization and mixed-precision training sounds mysterious, this’ll clear it up. 📺 youtu.be/Ue3AK4mCYYg

youtu.be

4-Bit Training for Billion-Parameter LLMs? Yes, Really.

YouTube video by AI Coffee Break with Letitia

120

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 23/03/2025

Just say “Wait…” – and your LLM gets smarter?! We explain how just 1,000 training examples + a tiny trick at inference time = o1-preview level reasoning. No RL, no massive data needed. 🎥 Watch now → youtu.be/XuH2QTAC5yI

youtu.be

s1: Simple test-time scaling: Just “wait…” + 1,000 training examples? | PAPER EXPLAINED

YouTube video by AI Coffee Break with Letitia

043

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 01/02/2025

It was wonderful reconnecting with Albert Gatt, @ecekt.bsky.social @annawegmann.bsky.social and for all new wonderful people I met, such as Antal van den Bosch, Catalina Goanta, and so many others! Thanks to the conference organisers for a wonderful conference! 🙌

020

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 01/02/2025

📍 Hosted at the stunning Railway Museum in Utrecht 🚂, and surrounded by the vibrant university community, I am grateful for the chance to share insights, engage with an inspiring community, and participate in stimulating research exchange at Utrecht University! 🌟

100

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 01/02/2025

🎙️ Yesterday, I gave a keynote on large language models outfitted with visual understanding, and the faithfulness of their chain-of-thought reasoning at the National Conference on Governing the Digital Society and Human-Centered AI.

120

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 01/02/2025

Thank you so much for the kind words of appreciation! 🤗

010

Reposted by aicoffeebreak.bsky.social

Robert A. Bagheri @ayoubbagheri.nl · 01/02/2025

The National Conference on AI Transformations: Language, Technology, and Society organised by Utrecht University @utrechtuniversity.bsky.social was a success, and indeed Letiția‘s @aicoffeebreak.bsky.social talk was very inspiring.

121

Reposted by aicoffeebreak.bsky.social

Heidelberg University NLP Group @hd-nlp.bsky.social · 27/01/2025

🎉 Exciting news from our team! The final paper of @aicoffeebreak.bsky.social's PhD journey is accepted at #ICLR2025! 🙌 🖼️📄 Check out her original post below for more details on Vision & Language Models (VLMs), their modality use and their self-consistency 🔥

092

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 26/01/2025

We explain 🥥COCONUT (Chain of Continuous Thought), a new paper using vectors for CoT instead of words. We break down: - Why CoT with words might not be optimal. - How to implement vectors for CoT instead words and make CoT faster. - What this means for interpretability. 📺 youtu.be/mhKC3Avqy2E

youtu.be

COCONUT: Training large language models to reason in a continuous latent space – Paper explained

YouTube video by AI Coffee Break with Letitia

010

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 22/01/2025

(3)🔎We provide an update of the accuracies reached by state-of-the-art VLMs on the VALSE 💃benchmark aclanthology.org/2022.acl-lon... 🎯 Even modern VLMs still struggle with most phenomena tested by VALSE💃, although there are strong improvements from models such as mplug-owl3.

aclanthology.org

VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena

Letitia Parcalabescu, Michele Cafagna, Lilitta Muradjan, Anette Frank, Iacer Calixto, Albert Gatt. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Lo...

100

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 22/01/2025

(2)🔎We evaluate VLMs' self-consistency when generating post-hoc and CoT explanations. 🎯 Most VLMs are less self-consistent than LLMs. For all models, the contributions of the image are significantly stronger when generating explanations compared to answers.

100

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 22/01/2025

(1)🔎We measure how much VLMs use text and images when generating predictions or explanations. 🎯 We find that VLMs are heavily text-centric when producing answers and natural language explanations.

100

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 22/01/2025

The last paper of my PhD is accepted at ICLR 2025! 🙌 🎊 We investigate the reliance of modern Vision & Language Models (VLMs) on image🖼️ vs. text📄 inputs when generating answers vs. explanations, revealing fascinating insights into their modality use and self-consistency. Takeaways: 👇

arxiv.org

Do Vision & Language Decoders use Images and Text equally? How Self-consistent are their Explanations?

Vision and language model (VLM) decoders are currently the best-performing architectures on multimodal tasks. Next to answers, they are able to produce natural language explanations, either in post-ho...

181

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 19/01/2025

An educational and a bit historical deep dive into LLM research. 💡Learn what breakthroughs since 2017 paved the way for AI like ChatGPT (it wasn't overnight). We go through: * Transformers * Prompting * Human Feedback, etc. and break it all down for you! 👇 📺 youtu.be/BprirYymXrg

youtu.be

LLMs Explained: A Deep Dive into Transformers, Prompts, and Human Feedback

YouTube video by AI Coffee Break with Letitia

031

Reposted by aicoffeebreak.bsky.social

Franz Nowak @franznowak.bsky.social · 18/11/2024

Transformer language models like Chat GPT, when using chain-of-thought reasoning, are Turing complete. Specifically, they can execute probabilistic algorithms and generate any computable weighted language youtu.be/MMIJKKNxvec?...

youtu.be

Transformer LLMs are Turing Complete after all !?

YouTube video by AI Coffee Break with Letitia

051

aicoffeebreak.bsky.social @aicoffeebreak.bsky.social · 08/12/2024

Thank you so much, Lenore! 💌 Love and greetings from Heidelberg!

000