Pruning as a Defense: Reducing Memorization in Large Language Models
Abstract
Large language models have been shown to memorize significant portions of their training data, which they can reproduce when appropriately prompted. This work investigates the impact of simple pruning techniques on this behavior. Our findings reveal that pruning effectively reduces the extent of memorization in LLMs, demonstrating its potential as a foundational approach for mitigating membership inference attacks.
Cite
Text
Gupta et al. "Pruning as a Defense: Reducing Memorization in Large Language Models." ICLR 2025 Workshops: BuildingTrust, 2025.Markdown
[Gupta et al. "Pruning as a Defense: Reducing Memorization in Large Language Models." ICLR 2025 Workshops: BuildingTrust, 2025.](https://mlanthology.org/iclrw/2025/gupta2025iclrw-pruning/)BibTeX
@inproceedings{gupta2025iclrw-pruning,
title = {{Pruning as a Defense: Reducing Memorization in Large Language Models}},
author = {Gupta, Mansi and Waghela, Nikhar and Gupta, Sarthak and Goel, Shourya and Shanmugavelu, Sanjif},
booktitle = {ICLR 2025 Workshops: BuildingTrust},
year = {2025},
url = {https://mlanthology.org/iclrw/2025/gupta2025iclrw-pruning/}
}