AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Alignment-Enhanced Integration of Connectivity and Spectral Sparse in Dynamic Sparse Training of LLM

Wenjing Wu, Yingtao Zhang, Jialin Zhao, Carlo Vittorio Cannistraci

ICLR 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Layer-wise pruning of LLMs

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 alignment loss on top of activation adjustment (ablation)CHTs + low-rank · OpenWebText LLaMA-60M, sparsity 0.9validation perplexity: 32.21 -> 31.77Improved
add activation adjustment to naive sum of CHTs and low-rank (ablation)CHTs + low-rank · C4 LLaMA-60M, sparsity 0.9validation perplexity: 189.55 -> 39.66Improved
add activation adjustment to naive sum (ablation)CHTs + low-rank · OpenWebText LLaMA-130M, sparsity 0.9validation perplexity: 119.35 -> 24.45Improved
add activation adjustment to naive sum (ablation)CHTs + low-rank · OpenWebText LLaMA-60M, sparsity 0.7validation perplexity: 27.89 -> 29.94Hurt
add activation adjustment to naive sum (ablation)CHTs + low-rank · OpenWebText LLaMA-60M, sparsity 0.8validation perplexity: 33.35 -> 29.42Improved
add full alignment scheme (activation + alignment loss) versus naive sum (ablation)CHTs + low-rank · all models/datasets/sparsities (Wilcoxon)validation perplexity: p-value 0.00049Improved
add alignment loss on top of activation adjustment (ablation)CHTs + low-rank · all models/datasets/sparsities (Wilcoxon)validation perplexity: p-value 0.00049Improved
add activation adjustment alone versus naive sum (ablation)CHTs + low-rank · all models/datasets/sparsities (Wilcoxon)validation perplexity: p-value 0.0093Improved
replace static sparse branch with dynamic CHTs branch plus alignmentSLTrain -> CHTsL · OpenWebText LLaMA-60M, sparsity 0.9validation perplexity: 33.90 -> 31.77Improved
add low-rank branch with alignment to CHTsCHTs -> CHTsL · OpenWebText LLaMA-60M, sparsity 0.9validation perplexity: 33.03 -> 31.77Improved
add activation adjustment to static sparse plus low-rank (ablation)static sparse + low-rank · OpenWebText LLaMA-130M, sparsity 0.7validation perplexity: 21.25 -> 41.51Hurt
add alignment loss on top of activation adjustment (ablation)static sparse + low-rank · C4 LLaMA-130M, sparsity 0.7validation perplexity: 26.52 -> 26.36Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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