【LLM安全】丸一口：Agent安全调研

Name: 【LLM安全】丸一口：Agent安全调研
Uploaded: 2026-03-24T10:00:00.000Z
Duration: 1 h 3 min 23 s
Channel: 丸一口
Description: 这次主要是对以下论文内容进行汇报： [1] (姚顺雨ICLR2023) ReAct: Synergizing Reasoning and Acting in Language Models [2] (arxiv25) Agentic AI Security: Threats, Defenses, Evaluation, and Open Challenges [3] (ACL24,CCF-A)BadAgent: Inserting and Activating Backdoor Attacks in LLM Agents PPT&思维导图链接：https://github.com/Van-Ec

2847

2026-03-24 18:00:00

未经作者授权，禁止转载

这次主要是对以下论文内容进行汇报： [1] (姚顺雨ICLR2023) ReAct: Synergizing Reasoning and Acting in Language Models [2] (arxiv25) Agentic AI Security: Threats, Defenses, Evaluation, and Open Challenges [3] (ACL24,CCF-A)BadAgent: Inserting and Activating Backdoor Attacks in LLM Agents PPT&思维导图链接：https://github.com/Van-Echo/OpenWanYikou/blob/main/04-Presentation