AuthorizationBench: LLM agents exceed authority in 19.2% of runs
Across 6,864 trajectories, violation rates ranged from 10.5% to 26.6% and barely changed with more reasoning.
Stories per day · 30 days
1 stories, newest first · one card per story
1 stories, newest first · one card per story
Across 6,864 trajectories, violation rates ranged from 10.5% to 26.6% and barely changed with more reasoning.
Today 1.6K readers · 2.4K views · 774 reads · all time 189K views