2 years ago|discuss
user: zh217
506 karma | created 9 years ago
recent submissions
GPT-4 Is Too Smart to Be Safe: Stealthy Chat with LLMs via Cipher
(llmcipherchat.github.io)
3 pts|2 years ago|1 comment
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
Dromedary: Principle-Driven Self-Alignment of Language Models
(mitibmdemos.draco.res.ibm.com)
1 pts|2 years ago|1 comment
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
Reinforcement learning is all you need, for next generation language models
(yuxili.substack.com)
5 pts|2 years ago|discuss
3 pts|2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss
2 years ago|discuss