research/problems/llm_sql/large/baseline_cache.json

{
  "baseline_hit_rate": 0.09871281929519464,
  "individual_rates": [
    0.1441122329122357,
    0.05331340567815356
  ],
  "num_datasets": 2
}