Merge branch 'EleutherAI:main' into main

lm_eval/tasks/bigbench/multiple_choice/temporal_sequences.yaml

 # Generated by utils.py
 dataset_name: temporal_sequences_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_temporal_sequences_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/tense.yaml

-# Generated by utils.py
-dataset_name: tense_zero_shot
-include: ../multiple_choice_template_yaml
-task: bigbench_tense_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/timedial.yaml

 # Generated by utils.py
 dataset_name: timedial_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_timedial_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/topical_chat.yaml

-# Generated by utils.py
-dataset_name: topical_chat_zero_shot
-include: ../multiple_choice_template_yaml
-task: bigbench_topical_chat_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/tracking_shuffled_objects.yaml

 # Generated by utils.py
 dataset_name: tracking_shuffled_objects_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_tracking_shuffled_objects_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/understanding_fables.yaml

 # Generated by utils.py
 dataset_name: understanding_fables_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_understanding_fables_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/undo_permutation.yaml

 # Generated by utils.py
 dataset_name: undo_permutation_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_undo_permutation_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/unit_conversion.yaml

 # Generated by utils.py
 dataset_name: unit_conversion_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_unit_conversion_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/unit_interpretation.yaml

 # Generated by utils.py
 dataset_name: unit_interpretation_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_unit_interpretation_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/unnatural_in_context_learning.yaml

-# Generated by utils.py
-dataset_name: unnatural_in_context_learning_zero_shot
-include: ../multiple_choice_template_yaml
-task: bigbench_unnatural_in_context_learning_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/vitaminc_fact_verification.yaml

 # Generated by utils.py
 dataset_name: vitaminc_fact_verification_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_vitaminc_fact_verification_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/what_is_the_tao.yaml

 # Generated by utils.py
 dataset_name: what_is_the_tao_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_what_is_the_tao_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/which_wiki_edit.yaml

 # Generated by utils.py
 dataset_name: which_wiki_edit_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_which_wiki_edit_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/winowhy.yaml

 # Generated by utils.py
 dataset_name: winowhy_zero_shot
-include: ../multiple_choice_template_yaml
+include: ../multiple_choice_template_a_yaml
 task: bigbench_winowhy_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/word_sorting.yaml

-# Generated by utils.py
-dataset_name: word_sorting_zero_shot
-include: ../multiple_choice_template_yaml
-task: bigbench_word_sorting_multiple_choice

lm_eval/tasks/bigbench/multiple_choice/word_unscrambling.yaml

-# Generated by utils.py
-dataset_name: word_unscrambling_zero_shot
-include: ../multiple_choice_template_yaml
-task: bigbench_word_unscrambling_multiple_choice

lm_eval/tasks/bigbench/multiple_choice_template_yaml → lm_eval/tasks/bigbench/multiple_choice_template_a_yaml

@@ -12,4 +12,4 @@ metric_list:
   - metric: acc
   # TODO: brier score and other metrics
 metadata:
-  version: 0.0
+  version: 1.0

lm_eval/tasks/bigbench/multiple_choice_template_b_yaml

+group: bigbench_multiple_choice
+dataset_path: hails/bigbench
+dataset_kwargs:
+  # num_shots: 0 # TODO: num of shots for `bigbench` HF dataset should be controlled through this, not through the typical methods
+  # subtask_name: null
+output_type: multiple_choice
+test_split: default
+doc_to_text: inputs
+doc_to_target: "{{multiple_choice_scores.index(1)}}"
+doc_to_choice: "{{multiple_choice_targets}}"
+metric_list:
+  - metric: acc
+  # TODO: brier score and other metrics
+metadata:
+  version: 1.0

lm_eval/tasks/bigbench/push_bigbench_dataset.py

@@ -8,6 +8,7 @@ Requires the installation of
 `pip install "bigbench @ https://storage.googleapis.com/public_research_data/bigbench/bigbench-0.0.1.tar.gz"`
 and is included so that the bigbench dependency can be avoided.
 """
+
 import bigbench.api.util as bb_utils
 import datasets
 from tqdm import tqdm

lm_eval/tasks/blimp/_blimp.yaml

+group: blimp
+task:
+  - "blimp_adjunct_island"
+  - "blimp_anaphor_gender_agreement"
+  - "blimp_anaphor_number_agreement"
+  - "blimp_animate_subject_passive"
+  - "blimp_animate_subject_trans"
+  - "blimp_causative"
+  - "blimp_complex_NP_island"
+  - "blimp_coordinate_structure_constraint_complex_left_branch"
+  - "blimp_coordinate_structure_constraint_object_extraction"
+  - "blimp_determiner_noun_agreement_1"
+  - "blimp_determiner_noun_agreement_2"
+  - "blimp_determiner_noun_agreement_irregular_1"
+  - "blimp_determiner_noun_agreement_irregular_2"
+  - "blimp_determiner_noun_agreement_with_adj_2"
+  - "blimp_determiner_noun_agreement_with_adj_irregular_1"
+  - "blimp_determiner_noun_agreement_with_adj_irregular_2"
+  - "blimp_determiner_noun_agreement_with_adjective_1"
+  - "blimp_distractor_agreement_relational_noun"
+  - "blimp_distractor_agreement_relative_clause"
+  - "blimp_drop_argument"
+  - "blimp_ellipsis_n_bar_1"
+  - "blimp_ellipsis_n_bar_2"
+  - "blimp_existential_there_object_raising"
+  - "blimp_existential_there_quantifiers_1"
+  - "blimp_existential_there_quantifiers_2"
+  - "blimp_existential_there_subject_raising"
+  - "blimp_expletive_it_object_raising"
+  - "blimp_inchoative"
+  - "blimp_intransitive"
+  - "blimp_irregular_past_participle_adjectives"
+  - "blimp_irregular_past_participle_verbs"
+  - "blimp_irregular_plural_subject_verb_agreement_1"
+  - "blimp_irregular_plural_subject_verb_agreement_2"
+  - "blimp_left_branch_island_echo_question"
+  - "blimp_left_branch_island_simple_question"
+  - "blimp_matrix_question_npi_licensor_present"
+  - "blimp_npi_present_1"
+  - "blimp_npi_present_2"
+  - "blimp_only_npi_licensor_present"
+  - "blimp_only_npi_scope"
+  - "blimp_passive_1"
+  - "blimp_passive_2"
+  - "blimp_principle_A_c_command"
+  - "blimp_principle_A_case_1"
+  - "blimp_principle_A_case_2"
+  - "blimp_principle_A_domain_1"
+  - "blimp_principle_A_domain_2"
+  - "blimp_principle_A_domain_3"
+  - "blimp_principle_A_reconstruction"
+  - "blimp_regular_plural_subject_verb_agreement_1"
+  - "blimp_regular_plural_subject_verb_agreement_2"
+  - "blimp_sentential_negation_npi_licensor_present"
+  - "blimp_sentential_negation_npi_scope"
+  - "blimp_sentential_subject_island"
+  - "blimp_superlative_quantifiers_1"
+  - "blimp_superlative_quantifiers_2"
+  - "blimp_tough_vs_raising_1"
+  - "blimp_tough_vs_raising_2"
+  - "blimp_transitive"
+  - "blimp_wh_island"
+  - "blimp_wh_questions_object_gap"
+  - "blimp_wh_questions_subject_gap"
+  - "blimp_wh_questions_subject_gap_long_distance"
+  - "blimp_wh_vs_that_no_gap"
+  - "blimp_wh_vs_that_no_gap_long_distance"
+  - "blimp_wh_vs_that_with_gap"
+  - "blimp_wh_vs_that_with_gap_long_distance"
+aggregate_metric_list:
+  - metric: acc
+    aggregation: mean
+    weight_by_size: False
+metadata:
+  version: 2.0