Add BHS benchmark (#3265)

* run linter * add acc_norm

Add BHS benchmark (#3265)
* run linter * add acc_norm
331288bb · James A. Michaelov · GitHub · 84aa9f95 · 331288bb · 331288bb
Unverified Commit 331288bb authored Sep 02, 2025 by James A. Michaelov Committed by GitHub Sep 02, 2025
8 changed files
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_D_AP_V_ni_AN.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_D_AP_V_ni_AN.yaml
+dataset_name: swahili-N_of_Poss_D_AP_V_ni_AN
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_D_AP_V_ni_AN
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_D_AP_ni_AN.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_D_AP_ni_AN.yaml
+dataset_name: swahili-N_of_Poss_D_AP_ni_AN
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_D_AP_ni_AN
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_D_A_V.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_D_A_V.yaml
+dataset_name: swahili-N_of_Poss_D_A_V
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_D_A_V
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_D_A_V1_V2.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_D_A_V1_V2.yaml
+dataset_name: swahili-N_of_Poss_D_A_V1_V2
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_D_A_V1_V2
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_D_V.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_D_V.yaml
+dataset_name: swahili-N_of_Poss_D_V
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_D_V
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_D_ni_A.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_D_ni_A.yaml
+dataset_name: swahili-N_of_Poss_D_ni_A
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_D_ni_A
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_V.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_V.yaml
+dataset_name: swahili-N_of_Poss_V
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_V
--- a/lm_eval/tasks/bhs/swahili-N_of_Poss_ni_A.yaml
+++ b/lm_eval/tasks/bhs/swahili-N_of_Poss_ni_A.yaml
+dataset_name: swahili-N_of_Poss_ni_A
+include: _template_yaml
+task: bhs__swahili__N_of_Poss_ni_A