AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

LongLLaDA: Unlocking Long Context Capabilities in Diffusion LLMs

Xiaoran Liu, Yuerong Song, Zhigeng Liu, Zengfeng Huang, Qipeng Guo, Ziwei He, Xipeng Qiu

AAAI 2026 · Large language models › Language, safety and interpretability · approach: LLM pruning and scaling data · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
apply NTK RoPE scaling (λ=4) to LLaDA at 8k LongBenchLLaDA-8B-Instruct · LongBench 8kAvg score: 36.8 -> 40.6Improved
apply NTK scaling (λ=4) to LLaDA at 4k LongBenchLLaDA-8B-Base · LongBench 4kAvg score: 34.1 -> 34.0No clear effect
raise NTK factor from λ=4 to λ=14 at 8k (ablation)LLaDA-8B-Base · RULER 8kAvg score: 84.7 -> 86.8Improved
raise NTK factor from λ=14 to λ=31 at 16k (ablation)LLaDA-8B-Base · RULER 16kAvg score: 72.0 -> 78.0Improved
use NTK extrapolation instead of direct extrapolation at 16kLLaDA-8B-Base · RULER 16kAvg score: 19.5 -> 44.1 (λ=4)Improved
set λ=31 (up to 24k) and λ=55 (32k) in NIAH (ablation)LongLLaDA · NIAHretrievalMixed
increase diffusion sampling steps from 1 or 4 to 8 or more (ablation)LLaDA-8B-Base · NIAHretrieval depth: fail beyond 8k at s=1,4; 25% depth at 16k and almost 10% at 24k for s>=8Improved
replace autoregressive LLaMA3 with diffusion LLaDA (bidirectional training) under direct extrapolation8B models · NIAH beyond pretrain lengthretrieval: 54.25 vs 51.71 overall score (Base)Improved
use dLLM instead of AR LLM on RULER aggregation tasksLLaDA-8B vs LLaMA3-8B · RULER 4k AGGAGG score: LLaDA-8B-Base 65.2 vs LLaMA3-8B-Base 98.1Hurt
use dLLM instead of AR LLM on RULER QA tasksLLaDA-8B-Instruct vs LLaMA3-8B-Instruct · RULER 4k QAQA score: 90.5 vs 68.5Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy