tag:
  - longbench
task: longbench_samsum
dataset_path: THUDM/LongBench
test_split: test
dataset_name: samsum
doc_to_text: 'Summarize the dialogue into a few short sentences. The following are some examples.\n\n{{context}}\n\n{{input}}'
doc_to_target: '{{answers}}'
generation_kwargs:
  max_gen_toks: 128
  temperature: 1
  do_sample: True
metric_list:
  - metric: !function metrics.rouge_score
    aggregation: mean
    higher_is_better: True
metadata:
  version: 1.0