Challenge: Existing approaches to prune LLMs rely on the C4 dataset as calibration data . arithmetic datasets perform better than pre-training datasets for pruning, whereas chain-of-thought is only useful on certain tasks.
Approach: They evaluate the selection of calibration data for LLM pruning across a wide range of datasets . they find that C4 is not the optimal calibration data, and that CoT is only useful on certain tasks.
Outcome: The chosen calibration data significantly impacts the performance of pruned LLMs, the authors found . their results shed light on the importance of carefully selecting calibration data for LLM pruning .

Similar Papers

On the Limitations of Language-targeted Pruning: Investigating the Calibration Language Impact in Multilingual LLM Pruning (2026.tacl-1)

Copied to clipboard

Challenge: Recent advances in large language model pruning have shown high predictive performance in post-training settings.
Approach: They conduct an empirical study on the performance and internal representation changes associated with pruning multilingual models for monolingual applications.
Outcome: The proposed pruning methods retain perplexity and yield high signal-to-noise ratios, but not consistently improve downstream tasks.
On the Impact of Calibration Data in Post-training Quantization and Pruning (2024.acl-long)

Copied to clipboard

Challenge: Quantization and pruning are the foundations of compression for large language models . however, no prior work has investigated how calibration data impacts performance of compression methods.
Approach: They propose an empirical study on the effect of calibration data on LLM performance.
Outcome: The proposed methods improve performance in a post-training setting.
DRPruning: Efficient Large Language Model Pruning through Distributionally Robust Optimization (2025.acl-long)

Copied to clipboard

Challenge: Structured pruning reduces model size but often causes uneven degradation across domains, leading to biased performance.
Approach: They propose a method that dynamically adjusts the data distribution during training to restore balanced performance across heterogeneous and multi-tasking data.
Outcome: Experiments in monolingual and multilingual settings show that the proposed method surpasses similarly sized models in pruning and continued pretraining over perplexity, downstream tasks, and instruction tuning.
Swift Cross-Dataset Pruning: Enhancing Fine-Tuning Efficiency in Natural Language Understanding (2025.coling-main)

Copied to clipboard

Challenge: Current approaches for fine-tuning datasets rely on expensive sample ranking processes . data set pruning aims to select a subset of a dataset for efficient model training .
Approach: They propose a method that uses TF-IDF embeddings with geometric median to rapidly evaluate sample importance.
Outcome: The proposed method significantly reduces training and storage costs while maintaining model effectiveness.
Pruning as a Domain-specific LLM Extractor (2024.findings-naacl)

Copied to clipboard

Challenge: Large Language Models (LLMs) have exhibited remarkable proficiency across a wide array of NLP tasks.
Approach: They propose a method for pruning large language models using general or task-specific weights to extract a compressed, task-agnostic LLM.
Outcome: The proposed method extracts a compressed, domain-specific, and task- agnostic LLM by identifying LLM weights that are pivotal for general capabilities, like linguistic capability and multi-task solving, and domain- specific knowledge.
Rethinking Pruning Large Language Models: Benefits and Pitfalls of Reconstruction Error Minimization (2024.emnlp-main)

Copied to clipboard

Challenge: minimizing reconstruction error is not always ideal and can overfit calibration data.
Approach: They propose a method to prune large language models by divide and conquer . they propose minimizing reconstruction error by more than 90% by using calibration data .
Outcome: The proposed pruning approach generates high reconstruction errors . the proposed technique reduces reconstruction error by more than 90% .
Bypass Back-propagation: Optimization-based Structural Pruning for Large Language Models via Policy Gradient (2025.acl-long)

Copied to clipboard

Challenge: Recent pruning methods rely on heuristically hand-crafted metrics, leading to suboptimal performance.
Approach: They propose a method that optimizes pruning masks by minimizing back-propagation . they learn an underlying Bernoulli distribution to sample binary pruning mask samples .
Outcome: The proposed method is able to support global and heterogeneous pruning without back-propagation.
Iterative Structured Pruning for Large Language Models with Multi-Domain Calibration (2026.eacl-industry)

Copied to clipboard

Challenge: Existing models with unstructured pruning often yield irregular sparsity patterns that necessitate specialized hardware or software support.
Approach: They propose a structured pruning framework that eliminates entire architectural components and maintains compatibility with standard hardware accelerators.
Outcome: The proposed model pruning framework achieves significant compression with minimal performance degradation on multiple models across diverse downstream tasks.
Task-aware Block Pruning with Output Distribution Signals for Large Language Models (2026.findings-eacl)

Copied to clipboard

Challenge: Existing methods to estimate block importance rely on representation similarity or computationally expensive sensitivity analyses to estimate task-aware model behavior.
Approach: They propose a novel approach that quantifies block-level uncertainty from the statistics of each block’s early-exited output distribution on a calibration dataset.
Outcome: Experiments show that the proposed approach preserves downstream task performance while reducing inference latency and computational cost.
IG-Pruning: Input-Guided Block Pruning for Large Language Models (2025.emnlp-main)

Copied to clipboard

Challenge: Existing methods for efficient inference rely on fixed block masks, which can lead to suboptimal performance.
Approach: They propose an input-aware block-wise pruning method that dynamically selects layer masks at inference time.
Outcome: The proposed method outperforms state-of-the-art static depth pruning methods . it is particularly suitable for resource-constrained deployment scenarios .

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations