Benchmark Almanac

The Learning Agency Lab

We track 4 benchmarks that The Learning Agency Lab made or helped make.

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

Benchmarks

Name Kind Released Best score Status
Automated Student Assessment Prize 2.0 Models score source-based essays from seven prompts on a six-point scale. The competition measures agreement with human grades using quadratic weighted kappa. Other professions Apr 2024 0.84 Not in cards
Feedback Prize: English Language Learning Models predict six writing trait scores for English learners. The competition averages the root mean squared error across the six traits. Other professions Aug 2022 0.59 Not in cards
Feedback Prize: Predicting Effective Arguments Models predict whether each labeled argument element is ineffective, adequate, or effective. The competition scores the probabilities with multiclass logarithmic loss. Other professions May 2022 0.61 Not in cards
Feedback Prize: Evaluating Student Writing Models locate seven kinds of argument elements in student essays. The competition averages F1 scores across types, using word overlap to match spans. Other professions Dec 2021 Not in cards