# LLaDA: A Diffusion-Based Language Model for Revisable Text Generation

DevFeed: [LLaDA: A Diffusion-Based Language Model for Revisable Text Generation](<https://devfeed.tech/articles/llada-llms-that-don-t-gaslight-you-33456.md>)

Original publisher: [Read original article](<https://timkellogg.me/blog/2025/02/17/diffusion>)

Published: 2025-02-17T00:00:00Z

Content type: article

Language: en

Sources: [Tim Kellogg](<https://devfeed.tech/sources/tim-kellogg.md>)

Topics: [diffusion-transformers](<https://devfeed.tech/topics/diffusion-transformers.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [AI Architecture](<https://devfeed.tech/topics/ai-architecture.md>)

Tags: [ai-architecture](<https://devfeed.tech/tags/ai-architecture.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [hallucinations](<https://devfeed.tech/tags/hallucinations.md>), [language](<https://devfeed.tech/tags/language.md>), [llms](<https://devfeed.tech/tags/llms.md>), [model](<https://devfeed.tech/tags/model.md>)

## AI overview

The article introduces LLaDA, a language model that uses diffusion rather than autoregressive next-token prediction. It explains how this approach can generate and revise text globally, potentially helping with hallucinations, reasoning loops, and coherence in structured writing such as contracts.

## Source excerpt

A new language model uses diffusion instead of next-token prediction. That means the text it can back out of a hallucination before it commits. This is a big win for areas like law & contracts, where global consistency is valued