AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

NaRA: Noise-Aware LoRA for Parameter-Efficient Fine-Tuning of Diffusion LLMs

Shuaidi Wang, Zhan Zhuang, HUANG Ruping, Yu Zhang

ICML 2026 · Large language models › Instruction and fine-tuning · approach: Calibration and knowledge in LLM tuning · 0 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
add noise-aware hypernetwork core matrix (NaRA vs LoRA)LoRA on LLaDA-8B-Instruct · Commonsense reasoningaverage accuracy: 80.18 -> 84.09Improved
NaRA vs LoRA on base modelLoRA on LLaDA-8B-Base · Commonsense reasoningaverage accuracy: 76.10 -> 83.83Improved
NaRA vs LoRALLaDA-8B-Instruct · Math reasoning (4 benchmarks avg)average accuracy: 80.07 -> 80.89Improved
NaRA vs LoRALLaDA-8B-Base · Math reasoning (4 benchmarks avg)average accuracy: 76.29 -> 77.02Improved
NaRA vs LoRALLaDA-8B-Instruct · Code generationpass@1 average: 38.32 -> 39.32Improved
use Fourier embedding instead of scalar noise input (ablation)NaRA · Math reasoning (4 benchmarks avg)average accuracy: 80.26 -> 80.89Improved
use Fourier embedding instead of MLP embedding (ablation)NaRA · Math reasoning (4 benchmarks avg)average accuracy: 80.39 -> 80.89Improved
share one hypernetwork instead of separate ones per module (ablation)NaRA · Math reasoning (4 benchmarks avg)average accuracy: 80.13 -> 80.89Improved
eta 0.1 instead of 1.0 (ablation)NaRA · Math reasoning (4 benchmarks avg)average accuracy: 80.39 -> 80.89Improved
eta 0.1 instead of 0.01 (ablation)NaRA · Math reasoning (4 benchmarks avg)average accuracy: 80.26 -> 80.89Improved
continuous hypernetwork instead of 4 separate noise-interval LoRAs (ablation)Multi-LoRA · Math reasoning (4 benchmarks avg)average accuracy: 74.83 -> 80.89Improved
increase LoRA rank from 32 to 40 (ablation)LoRA · Math reasoning (4 benchmarks avg)average accuracy: 80.07 -> 80.19No clear effect

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