• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
AI
Announcement

One-hot flow is claimed to reach 50% on TinyGSM, up from 4%, without self-conditioning

The author says the work explores continuous diffusion language model design, including how to embed tokens.

Brandon AmosBA
Oscar DavisOD
2 Sources, 1h ago, first seen 1h ago

TLDR

The author says continuous diffusion language models have gone from niche to contender in 2026, but their design recipe remains unsettled. They claim their new work takes one-hot flow from 4% to 50% on TinyGSM without self-conditioning and explores design choices, including how to embed tokens.

Combined views

2.6K

2 Sources, first seen 1h ago

80 likes5 comments35 saves55 reposts

Combined views

2.6K

2 Sources, first seen 1h ago

80 likes5 comments35 saves55 reposts
Image: @osclsd on X

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Featured Source

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

2 Sources

Oscar Davis@osclsdContinuous Diffusion LMs went from niche to contender this year, but the recipe is shaky: why self-conditioning? How to embed tokens? 🤨 Our new work takes your one-hot flow from 4% to 50% on TinyGSM without self-conditioning, and explores the design space of CDLMs. 🔥🚀 (1/6)1h
Brandon Amos@brandondamosRT @osclsd: Continuous Diffusion LMs went from niche to contender this year, but the recipe is shaky: why self-conditioning? How to embed t…1h
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    2 Sources

    Oscar Davis@osclsdContinuous Diffusion LMs went from niche to contender this year, but the recipe is shaky: why self-conditioning? How to embed tokens? 🤨 Our new work takes your one-hot flow from 4% to 50% on TinyGSM without self-conditioning, and explores the design space of CDLMs. 🔥🚀 (1/6)1h
    Brandon Amos@brandondamosRT @osclsd: Continuous Diffusion LMs went from niche to contender this year, but the recipe is shaky: why self-conditioning? How to embed t…1h
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet