This page is still being polished. If you have thoughts, please share them via the feedback form.
Data on this page is preliminary and may change. Please do not share or cite these figures publicly.
Techniques to remove, bound, or modify learned model capabilities post-training.
Also in Model
Methods that adjust model parameters by editing or light retraining are proposed to protect privacy during the predeployment stage
Machine unlearning methods involve light parameter retraining to help LLMs forget private information (Eldan & Russinovich, 2023; Chen & Yang, 2023; Yao et al., 2023). These methods focus on selectively removing specific knowledge from an LLM especially when the LLM has learned sensitive data during training. By performing light retraining, these methods allow LLMs to “unlearn” specific patterns related to sensitive data without the need for extensive retraining, effectively reducing the risk of privacy breaches while maintaining the overall performance of LLMs. Model editing methods (Wu et al., 2023b; Chen et al., 2024b; Wu et al., 2024a) involve identifying parameters that are related to private information in LLMs and reducing the probability of outputting private data by editing or replacing these parameters. These techniques are particularly useful for mitigating the risk of unintended information leakage, such as when a model inadvertently recalls sensitive details from its training data (Chen et al., 2024b). By editing the parameters of a model, these methods target specific areas where sensitive information is stored, adjusting them to ensure that the model no longer outputs such information. In some cases, this can involve fine-tuning certain parameters, while in others, it may include replacing them with non-sensitive alternatives to preserve the functionality of LLMs without compromising privacy (Wu et al., 2024a).
Reasoning
Design standards establish privacy-protecting principles governing model development before deployment.
Privacy protection
Value Misalignment
99.9 OtherValue Misalignment > Mitigating social bias
1 AI SystemValue Misalignment > Privacy protection
1 AI SystemValue Misalignment > Methods for mitigating toxicity
1 AI SystemValue Misalignment > Methods for mitigating LLM amorality
1 AI SystemRobustness to attack
1 AI SystemLarge Language Model Safety: A Holistic Survey
Shi, Dan; Shen, Tianhao; Huang, Yufei; Li, Zhigen; Leng, Yongqi; Jin, Renren; Liu, Chuang; Wu, Xinwei; Guo, Zishan; Yu, Linhao; Shi, Ling; Jiang, Bojian; Xiong, Deyi (2024)
The rapid development and deployment of large language models (LLMs) have introduced a new frontier in artificial intelligence, marked by unprecedented capabilities in natural language understanding and generation. However, the increasing integration of these models into critical applications raises substantial safety concerns, necessitating a thorough examination of their potential risks and associated mitigation strategies. This survey provides a comprehensive overview of the current landscape of LLM safety, covering four major categories: value misalignment, robustness to adversarial attacks, misuse, and autonomous AI risks. In addition to the comprehensive review of the mitigation methodologies and evaluation resources on these four aspects, we further explore four topics related to LLM safety: the safety implications of LLM agents, the role of interpretability in enhancing LLM safety, the technology roadmaps proposed and abided by a list of AI companies and institutes for LLM safety, and AI governance aimed at LLM safety with discussions on international cooperation, policy proposals, and prospective regulatory directions. Our findings underscore the necessity for a proactive, multifaceted approach to LLM safety, emphasizing the integration of technical solutions, ethical considerations, and robust governance frameworks. This survey is intended to serve as a foundational resource for academy researchers, industry practitioners, and policymakers, offering insights into the challenges and opportunities associated with the safe integration of LLMs into society. Ultimately, it seeks to contribute to the safe and beneficial development of LLMs, aligning with the overarching goal of harnessing AI for societal advancement and well-being. A curated list of related papers has been publicly available at https://github.com/tjunlp-lab/Awesome-LLM-Safety-Papers.
Build and Use Model
Training, fine-tuning, and integrating the AI model
Developer
Entity that creates, trains, or modifies the AI system
Manage
Prioritising, responding to, and mitigating AI risks
