Kyle Lo

@kylelostat

language models, mts @MicrosoftAI, prev co-lead of Olmo @allen_ai, @uwcse, he/him, kylelo.bsky.social🧋

Seattle, WA

kyleclo.com

Joined January 2019

1,511 Following

4,734 Followers

Replies Reposts Media

Pinned
Kyle Lo
@kylelostat
Dec 17, 2025
olmo 3 paper finally on arxiv 🫡 thx to our teammates esp folks who chased additional baselines thx to arxiv-latex-cleaner and overleaf feature for chasing latex bugs thx for all the helpful discussions after our Nov release, best part of open science is progressing together!
15
Kyle Lo
@kylelostat
Nov 4
congrats on great open model! happy got to meet @DZhang50 thru this project, impressed how well yall pulled it together
Ricursive Intelligence
@RicursiveAI
Nov 4
Today we release Ripple-65B Preview, our 65B dense LLM we pretrained from scratch. Ripple sets a new state-of-the-art for U.S. open-weight models in agentic coding tasks, surpassing significantly larger models including Muse Spark Open-816B, Beam-501B, Inkling-975B, and Nemotron
3
Kyle Lo
@kylelostat
Oct 5
I'm also at #colm2026, Mon-Thurs Supporting two papers from @allen_ai days 🐟Olmo Hybrid, 7B hybrid model 🐠Cracks in the Foundation, long context recipes Hoping to have fun chats w folks about scaling LM data & evals, ☺️ unbiased on pretrain, posttrain, all the 🚂s
GIF
1
Kyle Lo
@kylelostat
Sep 25
before leaving, we released our dolma 3.5 data onto HF wout any fanfare, so happy folks r training on it 😍 very cool data mixing problem at this scale & dataset heterogeneity, congratss
Will Held
@WilliamBarrHeld
Sep 23
There’s an enormous amount of open training data on Hugging Face. What does it take to make it work together 🤗? For Marin’s 535B run, we built on 25T tokens from 152 datasets with licenses permitting training. Here’s the work between downloading those and training a model 🧵
1
Kyle Lo
@kylelostat
Jul 9
happening now at Hall A #1809, mining the web for diverse, natural “how to do X” instructions!
GIF
Yapei Chang
@YapeiChang
Apr 30
How2Everything will appear in ICML 2026! See you in Korea 🫡 We mine the web's procedural knowledge to better evaluate & train LLMs to generate valid step-by-step instructions, read more at: 🔗 arxiv.org/pdf/2602.08808
1

Kyle Lo

Kyle Lo’s posts