VOOZH about

URL: https://huggingface.co/trl-lib/llama-7b-se-rl-peft/discussions/7

⇱ trl-lib/llama-7b-se-rl-peft · Updates to Model Card noting issues with generated answers

Updates to Model Card noting issues with generated answers

#7
by nlothian - opened

It is very easy to generate extremely unethical answers (eg plans for genocide) using this model. This should be noted.

Here's an example.. 🤢

👁 image.png

lvwerra changed pull request status to merged

Thanks a lot for reporting and updating the README!

· Sign up or log in to comment