Log inSign up
Image
user avatar
Benjamin Minixhofer
@bminixhofer
PhD Student @CambridgeLTL
Cambridge, England
bmin.ai
Joined August 2015
422
Following
1,239
Followers
RepliesMedia
  • Pinned
    user avatar
    Benjamin Minixhofer
    @bminixhofer
    Dec 15, 2025
    We are releasing Bolmo today! Bolmo is the best byte-level model so far. It comes close to and sometimes surpasses Olmo 3. Bolmo also performs competitively in terms of speed & is fully open. I was skeptical of byte-level models for a long time but I finally switched camps🧵
    Image
    Image
    Image
    user avatar
    Ai2
    @allen_ai
    Dec 15, 2025
    Introducing Bolmo, a new family of byte-level language models built by "byteifying" our open Olmo 3—and to our knowledge, the first fully open byte-level LM to match or surpass SOTA subword models across a wide range of tasks. 🧵
  • user avatar
    Benjamin Minixhofer
    @bminixhofer
    Mar 5
    It's great to see Olmo Hybrid out! Huge congrats @lambdaviking @YanhongLi2062 @tyleraromero @AnejSvete and everyone involved. We had just converged on the final Bolmo architecture with linear RNNs in the local models when this project began. My small contribution was
    user avatar
    Ai2
    @allen_ai
    Mar 5
    Introducing Olmo Hybrid, a 7B fully open model combining transformer and linear RNN layers. It decisively outperforms Olmo 3 7B across evals, w/ new theory & scaling experiments explaining why. 🧵
    Image
  • user avatar
    Benjamin Minixhofer
    @bminixhofer
    Jan 21
    New blog post (my first!): Four Ingredients for Successful Retrofitting. If you're GPU-poor but want to do architecture research near the frontier, retrofitting is your friend. I wrote up what I've learned so far about what makes it work. Link ⬇️
    Image