# Rethinking Pruning Large Language Models: Benefits and Pitfalls of Reconstruction Error Minimization

arXiv: https://arxiv.org/abs/2406.15524 · HTML: https://mlsanity.com/paper/2406.15524/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 11 of its 41 references that are in the map.
- First posted on arXiv 21 Jun 2024.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 11 of 11)

- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Code Llama: Open Foundation Models for Code](https://mlsanity.com/paper/2308.12950/) (2023)
- [Mixtral of Experts](https://mlsanity.com/paper/2401.04088/) (2024)
- [BloombergGPT: A Large Language Model for Finance](https://mlsanity.com/paper/2303.17564/) (2023)
- [Gemma: Open Models Based on Gemini Research and Technology](https://mlsanity.com/paper/2403.08295/) (2024)
- [LLM-QAT: Data-Free Quantization Aware Training for Large Language Models](https://mlsanity.com/paper/2305.17888/) (2023)
- [FinGPT: Open-Source Financial Large Language Models](https://mlsanity.com/paper/2306.06031/) (2023)
- [CBQ: Cross-Block Quantization for Large Language Models](https://mlsanity.com/paper/2312.07950/) (2023)
- [EBFT: Effective and Block-Wise Fine-Tuning for Sparse LLMs](https://mlsanity.com/paper/2402.12419/) (2024)

## Built on by (most cited first, 8 of 8)

- [Beware of Calibration Data for Pruning Large Language Models](https://mlsanity.com/paper/2410.17711/) (2024)
- [On the Limitations of Language Targeted Pruning: Investigating the Calibration Language Impact in Multilingual LLM Pruning](https://mlsanity.com/paper/2408.14398/) (2024)
- [SAFE: Finding Sparse and Flat Minima to Improve Pruning](https://mlsanity.com/paper/2506.06866/) (2025)
- [The Unseen Frontier: Pushing the Limits of LLM Sparsity with Surrogate-Free ADMM](https://mlsanity.com/paper/2510.01650/) (2025)
- [Think Before You Prune: Selective Self-Generated Calibration for Pruning Large Reasoning Models](https://mlsanity.com/paper/2511.18864/) (2025)
- [A Free Lunch in LLM Compression: Revisiting Retraining after Pruning](https://mlsanity.com/paper/2510.14444/) (2025)
- [Beyond Single Plots: A Benchmark for Question Answering on Multi-Charts](https://mlsanity.com/paper/2604.21344/) (2026)
- [Relative Kinetic Utility for Reasoning-Aware Structural Pruning in Large Language Models](https://mlsanity.com/paper/2605.09008/) (2026)

## Abstract (arXiv)

This work suggests fundamentally rethinking the current practice of pruning large language models (LLMs). The way it is done is by divide and conquer: split the model into submodels, sequentially prune them, and reconstruct predictions of the dense counterparts on small calibration data one at a time; the final model is obtained simply by putting the resulting sparse submodels together. While this approach enables pruning under memory constraints, it generates high reconstruction errors. In this work, we first present an array of reconstruction techniques that can significantly reduce this error by more than 90%. Unwittingly, however, we discover that minimizing reconstruction error is not always ideal and can overfit the given calibration data, resulting in rather increased language perplexity and poor performance at downstream tasks. We find out that a strategy of self-generating calibration data can mitigate this trade-off between reconstruction and generalization, suggesting new directions in the presence of both benefits and pitfalls of reconstruction for pruning LLMs.