AI Paper Atlas

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

PASER: Post-Training Data Selection for Efficient Pruned Large Language Model Recovery

Bowei He, Lihao Yin, Hui-Ling Zhen, Xiaokun Zhang, Mingxuan Yuan, Chen Ma

ICLR 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 5 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 semantic-structural clustering (S2RIC) (ablation)PASER · LLM-Pruner 25% LLaMA2-7Baverage reasoning: 59.67 -> 61.10Improved
add semantic-structural clustering (S2RIC) (ablation)PASER · SliceGPT 25% LLaMA2-7Baverage reasoning: 63.03 -> 64.31Improved
add capability degradation-aware selection (CDAIS) (ablation)PASER · LLM-Pruner 25% LLaMA2-7Baverage reasoning: 60.26 -> 61.10Improved
add capability degradation-aware selection (CDAIS) (ablation)PASER · SliceGPT 25% LLaMA2-7Baverage reasoning: 62.68 -> 64.31Improved
add negative tuning effects mitigation (NTEM) (ablation)PASER · LLM-Pruner 25% LLaMA2-7Baverage reasoning: 59.25 -> 61.10Improved
add negative tuning effects mitigation (NTEM) (ablation)PASER · SparseGPT 50% LLaMA2-7Baverage reasoning: 61.76 -> 62.78Improved
recover with full Alpaca data instead of no trainingLLM-Pruner 25% · LLaMA2-7B Alpacaaverage reasoning: 57.78 -> 35.64Hurt
select data with PASER instead of NuggetsLLM-Pruner 25% · LLaMA2-7B Alpacaaverage reasoning: 58.69 -> 61.10Improved
select data with PASER instead of NuggetsSliceGPT 25% · LLaMA2-7B Alpacaaverage reasoning: 59.39 -> 64.31Improved
select data with PASER instead of NuggetsSparseGPT 50% · LLaMA2-7B AlpacaWikiText2 / average: 60.03 -> 62.78Improved
select data with PASER instead of Nuggets on LLaMA2-70BLLM-Pruner 50% · LLaMA2-70Baverage reasoning: 67.73 -> 69.62Improved
select data with PASER instead of Nuggets on Baichuan2-7BLLM-Pruner 25% · Baichuan2-7Baverage reasoning: 57.84 -> 59.70Improved

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