# BAAI proposes a safety defense for diffusion LLMs during denoising

> Original title: Why do standard defenses struggle with Diffusion LLMs (DLMs)?

- Source: BAAI
- Published: 2026-10-29T13:26:00.000Z
- Bad Influence: https://badinfluence.news/items/y2hfl36jxkwl2kv7k5gs9q0sk
- Original: https://x.com/BAAIBeijing/status/2115797300169539684

## Summary

BAAI says standard defenses struggle with diffusion LLMs because benign tokens can become risk-inducing in intermediate denoising states. Its method uses a lightweight safety value model trained with Spine Tree Search and K-branch rollouts, judged by Qwen Guard, to predict final safety from partial states. It then computes necessity and sufficiency risk attribution in one pass and re-masks only high-risk tokens.

## Post

Why do standard defenses struggle with Diffusion LLMs (DLMs)?

Unlike left-to-right autoregressive models, DLMs denoise tokens iteratively with bidirectional dependencies. Even benign tokens ("might","by","to") can act as risk-inducing tokens in intermediate states, shifting the conditional distribution toward an unsafe final output.

Instead of costly block-level regeneration or full-model retraining, BAAI exploits the editability of DLM intermediate states via three key designs:

🔹Lightweight Safety Value Model: Trained via Spine Tree Search & K-branch Rollouts(with Qwen Guard judge) to predict final safety from partial denoising states.

🔹Zero-Backbone-Rerun Counterfactual Readout: Reuses cached token features via Masked Cross-Attention (Full, Exclusion, and Single-token views) to computeNecessity & Sufficiencyrisk attribution in one lightweight pass.

🔹Surgical Re-masking: Selectively re-masks only high-risk tokens while keeping safe context intact.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHVyyynyaYAAqQtI.jpg&mode=thumb&exp=1793491200&sig=c49bcd0c443f5e47)

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHVyyynya0AAezrl.jpg&mode=thumb&exp=1793491200&sig=362f5d41b9da0bc3)