bars
大模型安全笔记
search
circle-xmark
Ctrl
k
copy
Copy
chevron-down
LLM-Attack
BadLlama: cheaply removing safety fine-tuning from Llama 2-Chat 13B
Previous
Evil Geniuses: Delving into the Safety of LLM-based Agents
chevron-left
Next
SHADOW ALIGNMENT: THE EASE OF SUBVERTING SAFELY-ALIGNED LANGUAGE MODELS
chevron-right