AI Paper Atlas

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

Improving Generalization in LLM Structured Pruning via Function-Aware Neuron Grouping

Tao Yu, Yongqi An, Kuan Zhu, Guibo Zhu, Ming Tang, Jinqiao Wang

AAAI 2026 · Large language models › Language, safety and interpretability · approach: LLM pruning and scaling data · 1 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 adaptive sparsity allocation (ASA) to OBC baseline (ablation)OBC · 7-task avg, 30% sparsityavg zero-shot accuracy: 57.88 -> 59.17Improved
add shared neuron group retention (SNGR) (ablation)OBC + ASA · 7-task avg, 30% sparsityavg zero-shot accuracy: 59.17 -> 59.64Improved
add function-aware pruning (FAP) (ablation)OBC + ASA + SNGR · 7-task avg, 30% sparsityavg zero-shot accuracy: 59.64 -> 59.83Improved
add function-aware pruning (FAP) (ablation)OBC + ASA + SNGR · 7-task avg, 40% sparsityavg zero-shot accuracy: 54.38 -> 54.96Improved
function-based neuron grouping instead of random grouping (ablation)O-FANG · 7-task avg, 40% sparsityavg zero-shot accuracy: 53.94 -> 54.96Improved
function-based neuron grouping instead of random grouping (ablation)O-FANG · 7-task avg, 30% sparsityavg zero-shot accuracy: 59.20 -> 59.83Improved
semantic-distance softmax reweighting instead of uniform weights (ablation)O-FANG · 7-task avg, 30% sparsityavg zero-shot accuracy: 59.47 -> 59.83Improved
semantic-distance reweighting instead of Only-Matched cluster (ablation)O-FANG · 7-task avg, 30% sparsityavg zero-shot accuracy: 55.90 -> 59.83Improved
semantic-distance reweighting instead of Reverse weights (ablation)O-FANG · 7-task avg, 40% sparsityavg zero-shot accuracy: 52.44 -> 54.96Improved
functional-complexity sparsity allocation instead of Taylor-based (ablation)O-FANG · LLaMA2-7B 30% sparsityavg zero-shot accuracy: 58.85 -> 59.83Improved
functional-complexity sparsity allocation instead of Taylor-based (ablation)O-FANG · LLaMA2-13B 40% sparsityavg zero-shot accuracy: 52.17 -> 61.16Improved
WikiText calibration with FANG instead of Alpaca calibration with OBC+ASAO-FANG · LLaMA2-7B 30%perplexity: 10.24 -> 7.23Improved

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