Benchmark Almanac

partner institutions

We track 3 benchmarks that partner institutions made or helped make.

0 in use 2 with public code 1 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
NewsBench (Chinese journalism) Models complete Chinese newsroom assignments and answer multiple choice questions about writing and editorial safety. GPT-4 evaluates the written responses. Other professions Feb 2024 2.44 Not in cards
Chinese Student English Essay Models score English essays from Beijing high school exams. Teachers assign up to 8 points for content, 8 for language, and 4 for structure. Other professions Jan 2024 Not in cards
MEEN 361 Conceptual Question Grading Models use an instructor’s rubrics to grade answers to mechanical engineering laboratory quizzes. Gao and colleagues compared model grades with human grades across ten questions. Other professions Nov 2024 Not in cards