{
  "appendix_count_status": "not_in_appendix",
  "appendix_n": null,
  "checkpoint": "main",
  "examples": 184,
  "examples_file": "examples/compositional__capital_upper.jsonl",
  "generation_settings": {
    "backend": "transformers",
    "max_tokens": 50,
    "temperature": 0.0,
    "top_p": 1.0
  },
  "historical_scorer_revision": null,
  "license": "CC-BY-4.0",
  "metrics_source": "results/olmo2_continuous_13b_final_iteration/OLMo-2-1124-13B_main/allenai_OLMo-2-1124-13B_main_compositional_capital_upper_metrics.json",
  "operations": [
    "country_to_capital+uppercase"
  ],
  "reference_model": "olmo2_13b",
  "saved_metric_names": [
    "continuous",
    "exact_match"
  ],
  "snapshot_policy": "one archived reference run; full prompts retained verbatim",
  "source_file": "results/olmo2_continuous_13b_final_iteration/OLMo-2-1124-13B_main/allenai_OLMo-2-1124-13B_main_compositional_capital_upper_capital_upper_detailed.jsonl",
  "source_module": "sources/compositional_task.py",
  "source_sha256": "8e976158f7b19c8a9ae5c461467f31b63fdd3896301e6bcb695d7973633e3887",
  "task_config": {
    "description": "Compositional task with chained string operations",
    "evaluation_metrics": [
      "accuracy"
    ],
    "input_column": "input",
    "max_new_tokens": null,
    "num_demonstrations": 5,
    "output_column": "output",
    "prompt_template": null
  },
  "task_id": "compositional:capital_upper",
  "task_version": "0.1"
}