AI Paper Atlas

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

ParallelBench: Understanding the Trade-offs of Parallel Decoding in Diffusion LLMs

Wonjun Kang, Kevin Galim, Seunghyuk Oh, Minjae Lee, Yuchen Zeng, Shuibai Zhang, Coleman Richard Charles Hooper, Yuezhou Hu, Hyung Il Koo, Nam Ik Cho, Kangwook Lee

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Parallel and early-exit decoding · 39 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
fine-tune LLaDA 1.5 on Waiting Line, then decode Replace Index in parallelLLaDA 1.5 · Waiting Line Replace Indexaccuracy: below 50% -> nearly 100%Improved
fine-tune LLaDA 1.5 and decode Replace Random and Shuffle in parallelLLaDA 1.5 · Waiting Line Replace Random / Shuffleaccuracy under parallel decodingNo clear effect
add Chain-of-Thought promptingLLaDA 1.5 · Waiting Lineaccuracy under parallelism: 8x more output tokensMixed
use Confidence Threshold instead of Top-k at conservative thresholdsLLaDA 1.5 · ParallelBenchspeed-quality trade-offImproved
use Confidence Threshold with aggressive thresholdsLLaDA 1.5 · ParallelBenchaccuracyHurt
use Confidence Top-k instead of Random Top-k on tasks with C(Y|X)=0 (Reverse, Replace Index)LLaDA 1.5 · Waiting LineaccuracyImproved
use Confidence Top-k instead of Random Top-k on tasks with C(Y|X)>0 (Replace Random, Shuffle)LLaDA 1.5 · Waiting LineaccuracyHurt
use small semi-AR blocks (more left-to-right)LLaDA 1.5 (Random Top-2) · Text Writing vs Waiting LineaccuracyMixed
use oracle per-sample threshold instead of fixed thresholdConfidence Threshold · ParallelBenchaccuracy and speedupImproved
increase parallelism k from 1 to larger on ShuffleLLaDA 1.5 · Shuffleaccuracy: near-perfect -> zeroHurt

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