Add the MultiBLiMP benchmark (#3155)

* add multiblimp * run linter

Add the MultiBLiMP benchmark (#3155)
* add multiblimp * run linter
de4ce482 · James A. Michaelov · GitHub · 8c025369 · de4ce482 · de4ce482
Unverified Commit de4ce482 authored Jul 19, 2025 by James A. Michaelov Committed by GitHub Jul 19, 2025
3 changed files
--- a/lm_eval/tasks/multiblimp/multiblimp_xnr.yaml
+++ b/lm_eval/tasks/multiblimp/multiblimp_xnr.yaml
--- a/lm_eval/tasks/multiblimp/multiblimp_xpg.yaml
+++ b/lm_eval/tasks/multiblimp/multiblimp_xpg.yaml
--- a/lm_eval/tasks/multiblimp/multiblimp_yrl.yaml
+++ b/lm_eval/tasks/multiblimp/multiblimp_yrl.yaml