src/backend/tasks/halueval/halueval_dialogue.yaml

task: halueval_qa
dataset_path: pminervini/HaluEval
dataset_name: dialogue_samples
output_type: generate_until
training_split: data
validation_split: data
test_split: data
num_fewshot: 0
doc_to_text: !function utils.doc_to_text_dialogue
doc_to_target: !function utils.doc_to_target_qa
process_results: !function utils.process_results_qa
metric_list:
  - metric: em
    aggregation: mean
    higher_is_better: true
  - metric: correctness
    aggregation: mean
    higher_is_better: true
metadata:
  - version: 0.0