请问能否介绍一下如何实现prompt攻击检测？ #1466
Unanswered
chengq2020
asked this question in
Q&A
请问能否介绍一下如何实现prompt攻击检测？
#1466
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
{{ message }}
-
有些用户可能会输入涉及歧视、黄暴、侵权等有安全合规风险隐患的对话提示词，常见的手法为目标劫持、角色扮演等，但我测试了几个问题 ChatGLM 都防范的很优秀，想请教一下实现方法和防御技巧，谢谢
Beta Was this translation helpful? Give feedback.
All reactions