Machine Unlearning of Pre-trained Large Language Models
It addresses ethical AI practices for users and developers by providing a framework for efficient unlearning in pre-trained models, though it is incremental as it builds on existing unlearning concepts.
This study tackles the problem of enabling the 'right to be forgotten' in pre-trained large language models by investigating machine unlearning methods, demonstrating that these methods are over 100,000 times more computationally efficient than retraining and improving hyperparameter robustness through gradient integration.
This study investigates the concept of the `right to be forgotten' within the context of large language models (LLMs). We explore machine unlearning as a pivotal solution, with a focus on pre-trained models--a notably under-researched area. Our research delineates a comprehensive framework for machine unlearning in pre-trained LLMs, encompassing a critical analysis of seven diverse unlearning methods. Through rigorous evaluation using curated datasets from arXiv, books, and GitHub, we establish a robust benchmark for unlearning performance, demonstrating that these methods are over $10^5$ times more computationally efficient than retraining. Our results show that integrating gradient ascent with gradient descent on in-distribution data improves hyperparameter robustness. We also provide detailed guidelines for efficient hyperparameter tuning in the unlearning process. Our findings advance the discourse on ethical AI practices, offering substantive insights into the mechanics of machine unlearning for pre-trained LLMs and underscoring the potential for responsible AI development.