lm_eval/tasks/super_glue/cb/default.yaml

tag:
  - super-glue-lm-eval-v1
task: cb
dataset_path: super_glue
dataset_name: cb
output_type: multiple_choice
training_split: train
validation_split: validation
doc_to_text: "{{premise}}\nQuestion: {{hypothesis}}. True, False, or Neither?\nAnswer:"
doc_to_target: label
doc_to_choice: ['True', 'False', 'Neither']
metric_list:
  - metric: acc
  - metric: f1
    aggregation: !function "aggregate.cb_multi_fi"
metadata:
  version: 1.0