Show-up attack study finds frontier LLM supervisors can be steered
On 188 SUA-Arena cases, all six tested frontier LLMs were vulnerable; GPT-5.6-Sol reached an ASR of 0.93.
Stories per day · 30 days
1 stories, newest first · one card per story
1 stories, newest first · one card per story
On 188 SUA-Arena cases, all six tested frontier LLMs were vulnerable; GPT-5.6-Sol reached an ASR of 0.93.
Today 520 readers · 576 views · 80 reads · all time 109K views