CryptoAILab/Awesome-LM-SSP

GitHub: CryptoAILab/Awesome-LM-SSP

系统收录大模型安全、安全性与隐私领域论文、工具、竞赛及排行榜的学术资源导航库。

Stars: 1884 | Forks: 126

# Awesome-LM-SSP [![Awesome](https://awesome.re/badge.svg)](https://awesome.re) [![Stars](https://img.shields.io/github/stars/ThuCCSLab/Awesome-LM-SSP)](.) [Awesome-LM-SSP](.) ## 介绍 与大模型 可信度相关的资源,涵盖多个维度(例如安全性、安全 和隐私),特别关注多模态 LM(例如视觉-语言模型 和扩散模型)。 - 本仓库正在更新中 :seedling: (手动收集)。 - 标签: - 模型: - ![LLM](https://img.shields.io/badge/LLM_(Large_Language_Model)-589cf4) - ![VLM](https://img.shields.io/badge/VLM_(Vision_Language_Model)-c7688b) - ![SLM](https://img.shields.io/badge/SLM_(Speech_Language_Model)-39c5bb) - ![Diffusion](https://img.shields.io/badge/Diffusion-a99cf4) - 备注:![Benchmark](https://img.shields.io/badge/Benchmark-87b800) ![New_dataset](https://img.shields.io/badge/New_dataset-87b800) ![Agent](https://img.shields.io/badge/Agent-87b800) ![CodeGen](https://img.shields.io/badge/CodeGen-87b800) ![Defense](https://img.shields.io/badge/Defense-87b800) ![RAG](https://img.shields.io/badge/RAG-87b800) ![Chinese](https://img.shields.io/badge/Chinese-87b800) ... - 来源:![conference](https://img.shields.io/badge/conference-f1b800) ![blog](https://img.shields.io/badge/blog-f1b800) ![OpenAI](https://img.shields.io/badge/OpenAI-f1b800) ![Meta AI](https://img.shields.io/badge/Meta_AI-f1b800) ... - 🔥🔥🔥 帮助我们更新列表! 🔥🔥🔥 - 首先,通过我们的数据库查看论文:[LM-SSP 元数据](https://docs.google.com/spreadsheets/d/1i2IfQJiAdFJueoy7sTv7snn__ZJx11GfiJx8rhDyfc0/edit?usp=sharing)。 - 如果你想更新论文信息(例如,一篇 arXiv 论文已被某个会议接收),请在我们的[元数据表](https://docs.google.com/spreadsheets/d/1i2IfQJiAdFJueoy7sTv7snn__ZJx11GfiJx8rhDyfc0/edit?usp=sharing)中搜索论文标题,然后在表格相应的单元格中留言。 - 如果你想添加论文,请通过 `ISSUE` 填写下表: | 标题 | 链接 | 代码 | 来源 | 分类 | 模型 | 备注 | | ---- |---- |---- |---- |---- |----|----| | 这是一个标题 | paper.com | github | bb'23 | A1. 越狱 | LLM | Agent | ## 新闻 - [2026.01.09] 🎂🎂 祝 Awesome-LM-SSP 两岁生日快乐!继续加油!💪 - [2025.01.09] 🎂 祝 Awesome-LM-SSP 一岁生日快乐!继续加油!💪 - [2024.01.09] 🚀 LM-SSP 已发布! ## 合集 - [书籍](collection/book.md) (3) - [竞赛](collection/competition.md) (5) - [排行榜](collection/leaderboard.md) (5) - [工具包](collection/toolkit.md) (21) - [综述](collection/survey.md) (40) - 论文 (2380) - A. Safety (安全) (1191) - [A0. 概述](collection/paper/safety/general.md) (30) - [A1. 越狱](collection/paper/safety/jailbreak.md) (532) - [A2. 对齐](collection/paper/safety/alignment.md) (147) - [A3. 深度伪造](collection/paper/safety/deepfake.md) (94) - [A4. 伦理](collection/paper/safety/ethics.md) (8) - [A5. 公平性](collection/paper/safety/fairness.md) (60) - [A6. 幻觉](collection/paper/safety/hallucination.md) (116) - [A7. 提示词注入](collection/paper/safety/prompt_injection.md) (118) - [A8. 毒性](collection/paper/safety/toxicity.md) (86) - B. Security (安全) (468) - [B0. 概述](collection/paper/security/general.md) (16) - [B1. 对抗样本](collection/paper/security/adversarial_examples.md) (105) - [B2. Agent](collection/paper/security/agent.md) (137) - [B3. 投毒与后门](collection/paper/security/poison_&_backdoor.md) (183) - [B4. 侧信道](collection/paper/security/side-channel.md) (2) - [B5. 系统](collection/paper/security/system.md) (25) - C. Privacy (隐私) (721) - [C0. 概述](collection/paper/privacy/general.md) (55) - [C1. 污染](collection/paper/privacy/contamination.md) (17) - [C2. 数据重构](collection/paper/privacy/data_reconstruction.md) (64) - [C3. 成员推理攻击](collection/paper/privacy/membership_inference_attacks.md) (68) - [C4. 模型提取](collection/paper/privacy/model_extraction.md) (14) - [C5. 隐私保护计算](collection/paper/privacy/privacy-preserving_computation.md) (133) - [C6. 属性推理攻击](collection/paper/privacy/property_inference_attacks.md) (8) - [C7. 侧信道](collection/paper/privacy/side-channel.md) (10) - [C8. 遗忘](collection/paper/privacy/unlearning.md) (70) - [C9. 水印与版权](collection/paper/privacy/watermark_&_copyright.md) (282) ## 对社区的热爱 —— 谢谢你们! 🙏 [![Star History Chart](https://api.star-history.com/svg?repos=CryptoAILab/Awesome-LM-SSP&type=Date)](https://star-history.com/#CryptoAILab/Awesome-LM-SSP&Date) ## 致谢 - 组织者:[Tianshuo Cong (丛天硕)](https://tianshuocong.github.io/),[Xinlei He (何新磊)](https://xinleihe.github.io/),[Zhengyu Zhao (赵正宇)](https://zhengyuzhao.github.io/),[Yugeng Liu (刘禹更)](https://liu.ai/),[Delong Ran (冉德龙)](https://github.com/eggry) - 本项目受到了 [LLM Security](https://llmsecurity.net/),[Awesome LLM Security](https://github.com/corca-ai/awesome-llm-security),[LLM Security & Privacy](https://github.com/chawins/llm-sp), [UR2-LLMs](https://github.com/jxzhangjhu/Awesome-LLM-Uncertainty-Reliability-Robustness),[PLMpapers](https://github.com/thunlp/PLMpapers),[EvaluationPapers4ChatGPT](https://github.com/THU-KEG/EvaluationPapers4ChatGPT) 的启发

标签:AI治理, Apex, 人工智能安全, 可信赖AI, 合规性, 多模态模型, 大模型安全, 大模型隐私, 学术资源, 对抗攻击, 扩散模型, 敏感信息检测, 数据隐私, 机器学习, 深度学习, 网络安全, 视觉语言模型, 论文列表, 资源汇总, 防御加固, 防御机制, 隐私保护