Research map / Large language models / Language, safety and interpretability
GPTailor: Large Language Model Pruning Through Layer Cutting and Stitching
Guinan Su, Li Shen, Lu Yin, Shiwei Liu, Yanwu Yang, Jonas Geiping
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 |
|---|---|---|---|
| add layer selection from fine-tuned variants and layer merging to layer removal (ablation) | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks: 44.83 -> 48.55 | Improved |
| add layer merging on top of layer selection and removal (ablation) | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks: 43.20 -> 48.55 | Improved |
| add layer selection without merging to layer removal (ablation) | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks: 44.83 -> 43.20 | Hurt |
| merge layers across fine-tuned variants instead of within-model folding (ablation) | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks: 46.26 -> 48.55 | Improved |
| use multi-task calibration objective instead of MMLU-only (ablation) | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks: 45.62 -> 48.55 | Improved |
| use accuracy on multiple-choice calibration data instead of WikiText perplexity as search objective (ablation) | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks: 25.38 -> 48.55 | Improved |
| prune a family of fine-tuned variants instead of a single model | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks: 42.24 (ShortGPT) -> 44.83 (LR-only) -> 48.55 | Improved |
| apply GPTailor to Llama-3 8B instead of ShortGPT | GPTailor · OpenCompass 14 tasks (Llama-3-8B) | average accuracy retained vs dense: 62.79% -> 84.55% | Improved |
| prune at 50% ratio | GPTailor · OpenCompass 14 tasks (Llama-2-7B) | average accuracy over 14 benchmarks | No clear effect |
| use multi-fidelity SMAC with small calibration budgets | GPTailor · search efficiency | fraction of trials at full budget: only 22% of trials use full budget; 41.4% use minimum | Improved |
Closest papers by meaning
- Layer as Puzzle Pieces: Compressing Large Language Models through Layer Concatenation (NeurIPS 2025 · similarity 0.94)
- An Efficient Pruner for Large Language Model with Theoretical Guarantee (ICML 2025 · similarity 0.93)
- SlimLLM: Accurate Structured Pruning for Large Language Models (ICML 2025 · similarity 0.93)
- SlimGPT: Layer-wise Structured Pruning for Large Language Models (NeurIPS 2024 · similarity 0.93)
- Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning (ICLR 2024 · similarity 0.93)
- Fluctuation-Based Adaptive Structured Pruning for Large Language Models (AAAI 2024 · similarity 0.92)
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.
