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
Interpretability techniques address this challenge from the perspective of the model itself. They can serve as tools to determine whether LLMs have internalized specific knowledge and to eliminate private information through knowledge editing.
Firstly, by explaining the relationship between factual knowledge and neuron activation (Hase et al., 2023), we can investigate whether and where a piece of factual knowledge is stored within the model. Yin et al. (2024) have develop a semantically constrained projected gradient descent method to explore whether LLMs possess certain knowledge that is independent of input prompts. Stoehr et al. (2024) utilize high-gradient weights in shallow layer attention heads to precisely locate memorized passages within the model. This localization technique identifies specific attention heads, which are then fine-tuned to forget the memorized knowledge, thereby enhancing the privacy protection capability of LLMs.
Reasoning
Mitigation name lacks definition and evidence; insufficient information to identify focal activity or mechanism.
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
