The Fragility and Dynamics of LLM Pruning-Recovery Steps
cs.CL
Submitted: 2026-10-11
Updated: 2026-10-11
Code: https://github.com/tatsu-lab/stanford_alpaca
Terminology
Sources
- Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
- Training Verifiers to Solve Math Word Problems
- SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot
- JSONSchemaBench: A Rigorous Benchmark of Structured Outputs for Language Models
- The Llama 3 Herd of Models
- Measuring Massive Multitask Language Understanding
- Distilling the Knowledge in a Neural Network
- Beware of Calibration Data for Pruning Large Language Models
- Pruning Filters for Efficient ConvNets
- PAT: Pruning-Aware Tuning for Large Language Models
- LLM-Pruner: On the Structural Pruning of Large Language Models
- Pointer Sentinel Mixture Models
- Pruning Convolutional Neural Networks for Resource Efficient Inference
- WinoGrande: An Adversarial Winograd Schema Challenge at Scale
- Movement Pruning: Adaptive Sparsity by Fine-Tuning
- Language Models are Multilingual Chain-of-Thought Reasoners
- Gemma 3 Technical Report
- Self-Data Distillation for Recovering Quality in Pruned Large Language Models
- Finetuned Language Models Are Zero-Shot Learners
- Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering