Benchmark Almanac

Sapienza University of Rome

We track 3 benchmarks that Sapienza University of Rome made or helped make.

0 in use 3 with public code 1 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
Adversarial Humanities Benchmark Models respond to harmful requests rewritten in literary and scholarly styles. A judge measures how often the rewrites elicit unsafe answers. Jailbreaks and prompt injection Apr 2026 9.20% Not in cards
Assembly101-O Models watch people assemble toy vehicles and flag the first procedural mistake. Training uses correct procedures, and evaluation stops at the first mistake in an incorrect sequence. Video understanding Apr 2024 Not in cards
Epic-tent-O Models watch someone pitch a tent and flag procedural mistakes using only the video so far. The test includes 15 videos. Video understanding Apr 2024 Not in cards