Franck Lebeau @_Kcnarf

#AI in #NLP, dataviz expert, full stack dev, math_art and day-to-day enthusiast; also, PhD in CS, and trying to reduce my environmental footprint Joined December 2016

Tweets

4K
Followers

565
Following

413
Likes

4K

Rohan Paul @rohanpaul_ai

2 days ago

🚨 BAD news for Medical AI models. MASSIVE revelations from this @Microsoft paper. 🤯 Current medical AI models may look good on standard medical benchmarks but those scores do not mean the models can handle real medical reasoning. The key point is that many models pass tests…

174 833 4K 480K 3K

Download Image

Datawrapper @Datawrapper

4 days ago

In this week's Dispatch, you'll find data vis on politics, trains, and minerals, but also interactive tools to explore, and yet another data game at the end. 📊 🕹️ datawrapper.de/blog/data-vis-…

0 3 8 1K 2

Download Image

Amir Zur @AmirZur2000

2 months ago

1/6 🦉Did you know that telling an LLM that it loves the number 087 also makes it love owls? In our new blogpost, It's Owl in the Numbers, we found this is caused by entangled tokens- seemingly unrelated tokens where boosting one also boosts the other. owls.baulab.info

18 72 666 68K 472

tomaarsen @tomaarsen

2 months ago

😎 I just published Sentence Transformers v5.1.0, and it's a big one. 2x-3x speedups of SparseEncoder models via ONNX and/or OpenVINO backends, easier distillation data preparation with hard negatives mining, and more! See 🧵for the deets:

1 15 132 5K 43

Download Image

Antoine Chaffin @antoine_chaffin

2 months ago

Obviously it has been catched by @_reachsumit before the official announcement! 😁 I am very happy to announce that PyLate has now an associated paper and it has been accepted to CIKM! Very happy to share this milestone with my dear co-creator @raphaelsrty 🫶

Sumit @_reachsumit

2 months ago

0 7 47 6K 24

3 7 39 4K 9

Franck Lebeau @_Kcnarf

2 months ago

🤔Do you know that LLMs produce probabilities among each available token of the vocabulary. Only after comes the choice of the final outputed token. 👌Here is crystal clear, yet insightful, explanations of the various technics used to choose the next token

AI Coffee Break with Letitia @AICoffeeBreak

2 months ago

1 8 38 2K 23

Download Image

0 1 3 441 1

Florian Tramèr @florian_tramer

2 months ago

Are hallucinated references making it to arXiv? Yes, definitely! Since the release of Deep Research in February bogus references are on the rise (coincidence?) I wrote a blog post (link below) on my analysis (which hugely underestimates the true rate of hallucinations...)

9 27 288 34K 87

Download Image

Franck Lebeau @_Kcnarf

2 months ago

TLDR : 𝐀𝐈 + 𝐆𝐨𝐨𝐝 𝐕𝐒 𝐁𝐚𝐝 engineer Good engineer + AI ≥ 10* good Engineer ≥ 100* bad engineer + AI

Santiago @svpino

2 months ago

TLDR : 𝐀𝐈 + 𝐆𝐨𝐨𝐝 𝐕𝐒 𝐁𝐚𝐝 engineer Good engineer + AI ≥ 10* good Engineer ≥ 100* bad engineer + AI

184 180 2K 107K 275

0 0 0 27 0

Jo Kristian Bergum @jobergum

2 months ago

I think more AI builders now recognize that the core quality concern is context confusion, not context window length limitations. Lots of agent implementations now let users compress context to avoid quality degradation.

8 5 77 7K 16

Adam Rackis @AdamRackis

2 months ago

From the AI workshop I'm in: "The S in MCP stands for security"

46 181 2K 129K 171

Yung-Sung Chuang @YungSungChuang

2 months ago

Scaling CLIP on English-only data is outdated now… 🌍We built CLIP data curation pipeline for 300+ languages 🇬🇧We train MetaCLIP 2 without compromising English-task performance (it actually improves! 🥳It’s time to drop the language filter! 📝arxiv.org/abs/2507.22062 [1/5] 🧵

3 87 309 37K 130

Download Image

Franck Lebeau @_Kcnarf

2 months ago

Je plussoie

Steve Krouse @stevekrouse

2 months ago

Je plussoie

94 146 954 237K 483

Download Image

0 0 0 32 0

Datawrapper @Datawrapper

2 months ago

📊 In this week's Data Vis Dispatch: U.S. tariffs, humanitarian crisis in Gaza, and much more. 🗞️ datawrapper.de/blog/data-vis-…

0 1 5 937 1

Download Image

Franck Lebeau @_Kcnarf

2 months ago

Imho, this illustrates current 𝑨𝑰 𝒄𝒐𝒎𝒑𝒆𝒕𝒆𝒏𝒄𝒚, which is distinct from intelligence

Stated Clearly - Jon Perry @StatedClearly

2 months ago

Imho, this illustrates current 𝑨𝑰 𝒄𝒐𝒎𝒑𝒆𝒕𝒆𝒏𝒄𝒚, which is distinct from intelligence

97 250 4K 409K 894

Download Video

0 0 0 69 1

Etienne Jacob @etiennejcb

3 months ago

I’ve been exploring physarum-style simulations on and off for a long time, and I finally wrote an article sharing the techniques I’ve been using. It ranges from the classic physarum algorithm to some of my own weird tricks. I hope you enjoy it! bleuje.com/physarum-expla…

8 41 319 22K 218

Raphaël Sourty @raphaelsrty

3 months ago

With @LightOnIO we are thrilled to release pylate-rs 🚀⭐️ An efficient inference engine for late-interaction models written in Rust and based on Candle ⚡️ pylate-rs is the best Python library / Rust crate / NPM package to spawn late-interaction models in milliseconds.

5 17 102 27K 62

LightOn @LightOnIO

3 months ago

🏎️ Introducing PyLate-rs @raphaelsrty is back at it to make you love late-interaction models! After quickening the retrieval process with FastPlaid, he’s now making inference lightning-fast with this lightweight tool crafted in Rust for optimal speed and efficiency! 💫And…

Raphaël Sourty @raphaelsrty

3 months ago

5 17 102 27K 62

0 16 52 5K 28

Download Image

Antoine Chaffin @antoine_chaffin

3 months ago

I'll be covering Reason-ModernColBERT in tonight's presentation, so please come if you are interested! maven.com/p/1973fe/going… (And please be gentle, this is the first time I will be speaking live in front of this many people 😭)

Victoria Slocum @victorialslocum

3 months ago

11 90 502 37K 532

Download Image

5 14 87 6K 60

Victoria Slocum @victorialslocum

3 months ago

Looking for a cheaper, open source alternative to agentic RAG? Try multi-vector retrieval with Reason-ModernColBERT. Most search systems compare one big summary vector per document. Multi-vector retrieval is different - it keeps separate vectors for each word or phrase, then…