Heads of Security attention gates improve prompt-injection robustness
Authors report RL-Hammer attack success of 10% on HoS versus 97% on GPT-OSS-20B and 85% on SecAlign.
Stories per day · 30 days
1 stories, newest first · one card per story
1 stories, newest first · one card per story
Authors report RL-Hammer attack success of 10% on HoS versus 97% on GPT-OSS-20B and 85% on SecAlign.
Today 342 readers · 367 views · 287 reads · all time 158K views