AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Time Is a Feature: Exploiting Temporal Dynamics in Diffusion Language Models

Wen Wang, Bozhen Fang, Chenchen Jing, Yongliang Shen, Yangyi Shen, Qiuyu Wang, Hao Ouyang, Hao Chen, Chunhua Shen

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Masked diffusion language model decoding · 32 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 exponential-weighted temporal votingLLaDA-8B-Instruct · GSM8K (gen length 256)accuracy: 76.3 -> 78.7Improved
add exponential-weighted temporal votingLLaDA-8B-Instruct · Countdown (gen length 128)accuracy: 20.3 -> 25.0 (+4.7)Improved
use fixed weighting instead of exponential in temporal voting (ablation)LLaDA-8B-Instruct · GSM8K (gen length 256)accuracy: 76.3 -> 73.4Hurt
use linear weighting for temporal voting (ablation)LLaDA-8B-Instruct · MATH500 (gen length 256)accuracy: 33.4 -> 34.4Improved
vary exponential alpha from 1 to 11 in temporal voting (ablation)LLaDA-8B-Instruct · four math benchmarksaverage accuracy gain: peak average gain of 1.5% at alpha=5Improved
add negative TSE reward in RFT (no labels)LLaDA-8B-Instruct SFT · Countdown (gen length 256)accuracy: 19.9 -> 53.5Improved
replace accuracy reward (d1) with negative TSE rewardLLaDA-8B-Instruct SFT · Countdownaverage accuracy gain: 15.1% (d1) -> 24.7% (TSE)Improved
combine TSE reward with accuracy rewardLLaDA-8B-Instruct SFT · SVAMP (gen length 512)accuracy: 88.7 -> 92.3Improved
combine TSE reward with accuracy rewardLLaDA-8B-Instruct SFT · Countdown (gen length 512)accuracy: 21.5 -> 54.7Improved
add temporal voting after RFT with negative TSE rewardLLaDA-8B-Instruct RFT · GSM8K (gen length 128)accuracy: 72.2 -> 72.6Improved
add temporal voting to the Sudoku-evaluated modelLLaDA-8B-Instruct SFT · Sudoku (gen length 512)accuracy: 5.5 -> 2.8Hurt
use negative TSE reward alone in RFTLLaDA-8B-Instruct SFT · Sudoku (gen length 512)accuracy: 5.5 -> 3.3Hurt

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