tawer12

SecAlign-Llama-3.1-8B-injection-at-end-inner5

tawer12/SecAlign-Llama-3.1-8B-injection-at-end-inner5

SecAlign-Llama-3.1-8B-injection-at-end-inner5 is a text generation model from tawer12. Use it when you need the model to write or continue text. It is set up for peft.

LoRA adapter trained with an SFT continuous attacker placed immediately after the text-injection span. Uses up to 10 attack tokens and 5 attacker optimization steps.