lm-evaluation/lm_eval/tasks/lambada/lambada_openai.yaml

group:
  - lambada
task: lambada_openai
dataset_path: EleutherAI/lambada_openai
dataset_name: default
output_type: loglikelihood
test_split: test
doc_to_text: "{{text.split(' ')[:-1]|join(' ')}}"
doc_to_target: "{{' '+text.split(' ')[-1]}}"
should_decontaminate: true
doc_to_decontamination_query: "{{text}}"
metric_list:
  - metric: perplexity
    aggregation: perplexity
    higher_is_better: false
  - metric: acc
    aggregation: mean
    higher_is_better: true
metadata:
  version: 1.0
dataset_kwargs:
  trust_remote_code: true