Large Language Model Safety: A Holistic Survey
Dan Shi, Tianhao Shen, Yufei Huang, Zhigen Li, Yongqi Leng, Renren Jin, Chuang Liu, Xinwei Wu, Zishan Guo, Linhao Yu, Ling Shi, Bojian Jiang, Deyi Xiong · 2024
Read Large Language Model Safety: A Holistic Survey by Dan Shi, Tianhao Shen, Yufei Huang, Zhigen Li, Yongqi Leng, Renren Jin, Chuang Liu, Xinwei Wu, Zishan Guo, Linhao Yu, Ling Shi, Bojian Jiang, Deyi Xiong (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-17686.txt.
Description
Dan Shi and co-authors survey the safety of large language models from every side. Their 2024 paper organizes the risks into four families: values that drift from human intent, fragility under adversarial attack, deliberate misuse, and the hazards of autonomous AI.
For each family they review defenses and the resources used to test them. Four side topics get their own treatment. One asks what safety means once models act as agents. Another covers how interpretability can expose trouble early, plus the safety roadmaps major AI firms have published. The last takes up governance, from national policy ideas to international cooperation. The authors close with a call for a proactive blend of engineering, ethics, and rules, and maintain a public reading list alongside the survey.
Catalog record
| CREATOR | Dan Shi, Tianhao Shen, Yufei Huang, Zhigen Li, Yongqi Leng, Renren Jin, Chuang Liu, Xinwei Wu, Zishan Guo, Linhao Yu, Ling Shi, Bojian Jiang, Deyi Xiong |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 258 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-17686