AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Pruning as a Cooperative Game: Surrogate-Assisted Layer Contribution Estimation for Large Language Models

Xuan Ding, Pengyu Tong, Ranjie Duan, Yunjian Zhang, Rui Sun, Yao Zhu

ICLR 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Layer-wise pruning of LLMs · 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
prune after 4-bit quantization instead of before (3 layers) (ablation)Shapley layer pruning + GPTQ · WikiText2, LLaMA-2-7Bperplexity: 15.4001 -> 14.9263Improved
prune after 4-bit quantization instead of before (6 layers) (ablation)Shapley layer pruning + GPTQ · WikiText2, LLaMA-2-7Bperplexity: 19.7741 -> 18.5761Improved
split pruning around quantization (3 + 3 layers) vs pruning 6 layers first (ablation)Shapley layer pruning + GPTQ · WikiText2, LLaMA-2-7Bperplexity: 19.7741 -> 19.4676Improved
re-estimate with joint pair deletion (Optimal Combo) instead of static two least important (ablation)layer deletion · BookCorpus PPL, LLaMA-2-7Bperplexity: 15.4535 -> 15.4279Improved
use Shapley-surrogate pruning vs ShortGPT, remove 9 layersMeta-LLaMA-3-8B · WikiText2perplexity: 2549.7485 -> 45.2635Improved
use Shapley-surrogate pruning vs SLEB, remove 12 layersMeta-LLaMA-3-8B · WikiText2perplexity: 126.9445 -> 304.5240Hurt
use Shapley-surrogate pruning vs SLEB, remove 3 layersLLaMA-2-7B · WikiText2perplexity: 14.2428 -> 14.6949Hurt
use Shapley-surrogate pruning vs SLEB, remove 12 layersLLaMA-2-7B · WikiText2perplexity: 58.1194 -> 38.1157Improved
use Shapley-surrogate pruning vs Shortened-LLaMA, 6.1BLLaMA-2-7B · zero-shot averageaccuracy: 0.5816 -> 0.5782No clear effect
use Shapley-surrogate pruning vs SliceGPT, 5.5BLLaMA-2-7B · zero-shot averageaccuracy: 0.3865 -> 0.5227Improved
apply pruning to Mamba-2.8B instead of ShortGPTMamba-2.8B · WikiText2perplexity: 4074.4865 -> 31.1091Improved
test surrogate on masks differing from training masks (ks 16,14,12,21,18) (ablation)surrogate network · LLaMA-2-7B mask predictionR2: 0.9464 -> 0.2466Hurt

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