Thibault Formal

38 posts

Thibault Formal

@thibault_formal

mts @ linkup

Paris

Joined December 2025

RepliesReplies RepostsReposts MediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone

or

Log in with username or email

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.

Thibault Formal
@thibault_formal
Sep 17
Introducing SPARSEUP: the first model release from @Linkup_platform , and the missing sparse companion of DenseOn and LateOn. Same backbone, same data, <150M, 56+ on BEIR-13. The model is open-source, use it! Blog: linkup.so/blog/introduci… Model: huggingface.co/Linkup-Platfor… (Apache
Linkup - SPARSEUP - Open-Source Sparse Retrieval Model (SOTA on BEIR-13)
From linkup.so
11
Thibault Formal
@thibault_formal
Mar 25
New sparse retrieval model for code retrieval: @simon_lupart did a great job during his internship!
Simon Lupart
@simon_lupart
Mar 24
Most code search systems rely on dense embeddings. In this work, we release SPLADE-Code, learned sparse retrieval models for code retrieval, with strong generalization, high interpretability, compatibility with inverted indexes, and working across 20+ programming languages.
Thibault Formal
@thibault_formal
Mar 17
New sparse retrieval model: introducing SPLARE, which extends SPLADE by replacing the vocabulary head with pretrained SAEs! paper: arxiv.org/abs/2603.13277 (ICLR'26) also how we won the WSDM'26 Cup on multilingual retrieval: arxiv.org/abs/2602.20986 (model weights coming soon!)
arxiv.org
Learning Retrieval Models with Sparse Autoencoders
Sparse autoencoders (SAEs) provide a powerful mechanism for decomposing the dense representations produced by Large Language Models (LLMs) into interpretable latent features. We posit that SAEs...
1