adversarial
Everything on Ground Truth tagged “adversarial” — 1 item.
'Leaderboard Hacking' paper: votes, prompts and decoy models can lift an arena ranking by up to 45 places News
A November 10 preprint by Fleisig, Baumann, Hadfield-Menell and Hovy shows that arena-style leaderboards built on the Bradley-Terry model can be manipulated, with each attack moving a model 15 to 20 places and all three combined up to 45.