Benchmark Almanac

Yale NLP Lab

We track 2 benchmarks that Yale NLP Lab made or helped make.

0 in use 2 with public code 2 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
AbGen Models design ablation studies to test the contribution of a specified component. The benchmark compares the proposed experiments with expert annotations from 807 NLP papers. AI research tasks Jul 2025 4.11 Not in cards
OpenComputer Computer-use agents complete 1,000 desktop tasks across 33 real applications, including browsers, office tools, creative software, development environments and messaging apps. App-specific verifiers inspect the final application state, and the harness gives partial credit from the recorded trajectory. Desktop apps May 2026 68.3% Not in cards