#DiffusionLLM - Twitter Hashtag

2 months ago

Couldn't be prouder! The core idea behind DiffusionGemma is one we introduced in RCD. That's what makes this work at such block sizes. Accepted to ICML 2026. See you in Seoul. Read more in this thread: #DiffusionLLM #ICML2026

Yuezhou Hu @yuezhouhu

3 months ago

Very excited to see that the core idea of DiffusionGemma directly stems from our work, Residual Context Diffusion (arXiv:2601.22954)! Code- and architecture-level comparisons are attached. 🏆 RCD is accepted to ICML 2026! See you in Seoul! #DiffusionLLM #LLM #Reasoning #GenAI

yuezhouhu's tweet photo. Very excited to see that the core idea of DiffusionGemma directly stems from our work, Residual Context Diffusion (arXiv:2601.22954)! Code- and architecture-level comparisons are attached.

🏆 RCD is accepted to ICML 2026! See you in Seoul!

#DiffusionLLM #LLM #Reasoning #GenAI https://t.co/Msy3fXQXM9

5

37

9

17

7K

0

10

2

1

1K

Yuezhou Hu @yuezhouhu

3 months ago

Very excited to see that the core idea of DiffusionGemma directly stems from our work, Residual Context Diffusion (arXiv:2601.22954)! Code- and architecture-level comparisons are attached. 🏆 RCD is accepted to ICML 2026! See you in Seoul! #DiffusionLLM #LLM #Reasoning #GenAI

5

37

9

17

7K

DeadNeuron @DeadNeuronML

3 months ago

Open question: SimSD assumes draft-model token distributions align enough with the dLLM target for meaningful acceptance rates. As dLLMs scale and diverge from AR drafters, does this degrade gracefully — or catastrophically? #MachineLearning #AI #DiffusionLLM

0

3

0

10

Gaurav Sarkar :) @GauravSarkar99

6 months ago

Not so much but current algo gives 5x faster text generation for the llada model. But no good conference to submit my work. Please suggest something good 🙏 #Diffusionllm

GauravSarkar99's tweet photo. Not so much but current algo gives 5x faster text generation for the llada model.
But no good conference to submit my work.
Please suggest something good 🙏

#Diffusionllm https://t.co/8bC8QbvTsg

0

6

0

158

Yuezhou Hu @yuezhouhu

7 months ago

Take a look at Residual Context Diffusion (RCD): a simple idea to boost diffusion LLMs—stop wasting “remasked” tokens!!! https://t.co/uDhQsuiCgK (Example on AIME24. RCD increases parallelism by 4x while reaching the baseline's peak accuracy.) #DiffusionLLM #LLM #Reasoning #GenAI

3

202

36

122

39K

Top Tweets for #DiffusionLLM

Last Seen Hashtags on Sotwe

Trends for you

Most Popular Users