Preprint
Oct 2026
Claim Laundering: When Agent Decision Models Trust Records Over Reality
Under injected silent failures, EVIGATE improves task success by 18.7 points in live ALFWorld agents, 5.0 in WebShop, and 5.1 in a phone simulator, and it also outperforms random vetoes at the same rate.
Tina Khezresmaeilzadeh, Seyedarmin Azizi, Erfan Baghaei Potraghloo et al.
· 0 citations