Log inSign up
Subham Sahoo
793 posts
Image
user avatar
Subham Sahoo
@ssahoo_
Pioneering Diffusion LLMs | Team Lead @mbzuai - IFM | PhD @cornell
San Francisco, CA
s-sahoo.com
Joined June 2010
81
Following
3,589
Followers
  • Pinned
    user avatar
    Subham Sahoo
    @ssahoo_
    Oct 6, 2025
    🎓 Officially a doctor now 😊!!! As a first-gen college kid, this moment means the world to me. Grateful beyond words to all my mentors who’ve guided me along the way — from @GMartius who first introduced me to research back in 2017, to @volokuleshov who sparked my love for
    Image
    102K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Sep 14, 2025
    For a PhD, you need to be a romantic at some level. Your papers will get rejected. Your ideas will get scooped. All while you peers flourish. And yes--It will sting. 2023 was one such year for me. Yet I call it my golden year, because that’s when I truly fell in love with my
    user avatar
    Hoang
    @hwangnamd
    Sep 12, 2025
    Replying to @ssahoo_
    What do you think would make a good PhD candidate? What specific traits do you see in a smart/talented PhD? Would love to hear some feedbacks on your side
    91K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Jun 13, 2025
    🚨 “The Diffusion Duality” is out! @ICML2025 ⚡️ Few-step generation in discrete diffusion language models by exploiting the underlying Gaussian diffusion. 🦾Beats AR on 3/7 zero-shot likelihood benchmarks. 📄 Paper: arxiv.org/abs/2506.10892 💻 Code: github.com/s-sahoo/duo 🧠
    Image
    GIF
    144K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Sep 12, 2025
    As I wrap up my thesis, I can’t help but look back on the past year of working on Diffusion LLMs. People often ask me: why and how I got into this strange little world of discrete diffusion. I usually give the textbook answer: the kind you’d find in any random paper and make
    Image
    38K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Oct 30, 2025
    Overwhelmed by the number of Diffusion LLM papers? 🌊 Same here 😭 So I’m starting a Discrete Diffusion Reading Group (@diffusion_llms) with my favorite disciples @jdeschena and @zhihanyang_ ✨ We’ll cover everything—from theory to empirics, from language to molecules. Join
    Image
    30K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Nov 11, 2025
    ✨New beginnings: I’ve joined the Institute of Foundation Models @llm360, where I’ll be leading research on diffusion-LLMs. 🚨Goals > Design frontier diffusion-LLMs > Advance these algorithms through fundamental research ✌️About to go on a hiring frenzy, so stay tuned.
    21K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Jun 3, 2025
    🚨 [New paper alert] Esoteric Language Models (Eso-LMs) First Diffusion LM to support KV caching w/o compromising parallel generation. 🔥 Sets new SOTA on the sampling speed–quality Pareto frontier 🔥 🚀 65× faster than MDLM ⚡ 4× faster than Block Diffusion 📜 Paper:
    Image
    93K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Oct 27, 2025
    🔥 Rethinking Reasoning (with Diffusion LLMs) This work changes how you think about reasoning in LLMs. 🤯 Turns out: you don’t need the full chain-of-thought — only a small subset of CoT tokens actually matter for the final answer. ❌ Autoregressive LLMs can’t exploit this
    Image
    15K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Sep 7, 2025
    Pre-training for Diffusion LLMs will be solved in the next 6 months. ^That’s underestimating both myself and the community.
    31K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Jul 14, 2025
    Attending ICML ✈️Tues-Fri to present "The Diffusion Duality" 🗓️Wed, July 16 @ 4:30pm 📍East Exhibition Hall A-B (E-3003) DM if you want to chat about diffusion LMs, or my current work on Duality or Esoteric LMs!
    user avatar
    Subham Sahoo
    @ssahoo_
    Jun 13, 2025
    🚨 “The Diffusion Duality” is out! @ICML2025 ⚡️ Few-step generation in discrete diffusion language models by exploiting the underlying Gaussian diffusion. 🦾Beats AR on 3/7 zero-shot likelihood benchmarks. 📄 Paper: arxiv.org/abs/2506.10892 💻 Code: github.com/s-sahoo/duo 🧠
    Image
    GIF
    10K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Nov 2, 2025
    We’re building a space that connects researchers, students, and practitioners working on discrete diffusion. Join the Discord — collaborate, learn, and share! Whether you’re 💼hiring or showcasing your work, this is the place 👇 Discord:
    user avatar
    Discrete Diffusion Reading Group
    @diffusion_llms
    Nov 2, 2025
    The Discrete Diffusion Reading Group is growing — 400+ members strong! We’ve launched a Discord for discussions, research ideas, help, and job opportunities. Join the conversation 👇 💬 discord.gg/JxSCwpNb 📧 groups.google.com/g/diffusion-ll…
    Image
    Image
    Discord - Group Chat That’s All Fun & Games
    From discord.com
    16K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Oct 11, 2025
    We’re dropping “The Diffusion Duality, Chapter 2” soon! So, stay tuned 🤗
    user avatar
    Sander Dieleman
    @sedielem
    Oct 10, 2025
    In diffusion LMs, discrete methods have all but displaced continuous ones (🥲). Interesting new trend: why not both? Use continuous methods to make discrete diffusion better. Diffusion duality: arxiv.org/abs/2506.10892 CADD: arxiv.org/abs/2510.01329 CCDD: arxiv.org/abs/2510.03206
    11K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Nov 13, 2025
    🚨“We have only one internet” (@ilyasut) — and that’s exactly why diffusion is the future of LLMs. 🔥Come for the hot takes, stay for @mihirp98’s deep dive at Monday’s @diffusion_llms reading group. ⏲️10 am ET (4pm CET)
    Image
    5.2K
  • user avatar
    Subham Sahoo
    @ssahoo_
    Oct 16, 2025
    Impressive work by @jdeschena ! They propose to replace the Encoder only denoising transformer with an Encoder-Decoder architecture which leads to faster training and inference of MDLM.
    user avatar
    Justin Deschenaux
    @jdeschena
    Oct 16, 2025
    📢 « Partition Generative Modeling (PGM): Masked Modeling without Masks » is out! 🚯 Masked diffusion models waste FLOPs processing countless mask tokens that carry no real information. ⚡We show how partitioning can replace masking, boosting throughput by >5.3x on text and up
    Image
    7.5K

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up