Research map / Architectures and efficiency / Pruning and efficient tuning
The Emergence of Essential Sparsity in Large Pre-trained Models: The Weights that Matter
AJAY KUMAR JAISWAL, Shiwei Liu, Tianlong Chen, Zhangyang Wang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Prune lowest-magnitude weights one-shot before fine-tuning | bert-base · GLUE and SQuAD downstream tasks | fine-tuning performance: 21.50 | No clear effect |
| Prune OPT-1.3B beyond 40% sparsity vs bert-base beyond 60% | OPT-1.3B · RTE | fine-tuning performance: 40% vs 60% | Hurt |
| Use self-supervised DINO pre-training instead of supervised ViT | ViT-base architecture · CIFAR-10, CIFAR-100, Tiny-ImageNet | zero weights and robustness to pruning: 14.37 | Improved |
| Increase pre-training data volume from 60% to 100% of bookcorpus (ablation) | bert-base · MNLI | prunability and downstream accuracy | Improved |
| Replace expensive LTH masks with free OMP masks within essential sparsity range (ablation) | bert-base, ViT-base · multiple downstream tasks | fine-tuning performance: $>96\%$ | No clear effect |
| Use SparseGPT instead of OMP for pruning (ablation) | Vicuna-7B · MMLU | performance at high sparsity | Improved |
| Apply N:M structured pruning instead of unstructured (ablation) | OPT-350M · MNLI, QNLI, RTE, SST-2 | fine-tuning performance drop | No clear effect |
Closest papers by meaning
- Scaling Laws for Sparsely-Connected Foundation Models (ICLR 2024 · similarity 0.88)
- No Free Prune: Information-Theoretic Barriers to Pruning at Initialization (ICML 2024 · similarity 0.87)
- The Journey Matters: Average Parameter Count over Pre-training Unifies Sparse and Dense Scaling Laws (ICLR 2025 · similarity 0.87)
- Sparsing Law: Towards Large Language Models with Greater Activation Sparsity (ICML 2025 · similarity 0.87)
- DEEP NEURAL NETWORK INITIALIZATION WITH SPARSITY INDUCING ACTIVATIONS (ICLR 2024 · similarity 0.86)
- Effective Interplay between Sparsity and Quantization: From Theory to Practice (ICLR 2025 · similarity 0.86)
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.
