delivery_state.json

{
  "completion_rule": "All 130 checkpoint evaluations must pass the 378-prompt, n=160, 60,480-scored-row, 8-raw-chunk, 8-scored-chunk, immutable-lineage, offline-W&B, and evaluation_complete gates.",
  "coverage": {
    "llama3b_sft_rp_ce_lam1": 10,
    "llama3b_sft_rp_hard_lam1": 10,
    "llama3b_sft_rp_uniform_lam1": 10,
    "q25_sft_ordered": 10,
    "q25_sft_rp_ce_lam1": 10,
    "q25_sft_rp_hard_lam1": 10,
    "q25_sft_rp_uniform_lam1": 10,
    "q25_sft_shuffled": 10,
    "q3_1p7b_sft_ordered": 10,
    "q3_1p7b_sft_rp_ce_lam1": 10,
    "q3_1p7b_sft_rp_hard_lam1": 10,
    "q3_1p7b_sft_rp_uniform_lam1": 10,
    "q3_1p7b_sft_shuffled": 10
  },
  "expected_coverage": {
    "llama3b_sft_rp_ce_lam1": 10,
    "llama3b_sft_rp_hard_lam1": 10,
    "llama3b_sft_rp_uniform_lam1": 10,
    "q25_sft_ordered": 10,
    "q25_sft_rp_ce_lam1": 10,
    "q25_sft_rp_hard_lam1": 10,
    "q25_sft_rp_uniform_lam1": 10,
    "q25_sft_shuffled": 10,
    "q3_1p7b_sft_ordered": 10,
    "q3_1p7b_sft_rp_ce_lam1": 10,
    "q3_1p7b_sft_rp_hard_lam1": 10,
    "q3_1p7b_sft_rp_uniform_lam1": 10,
    "q3_1p7b_sft_shuffled": 10
  },
  "expected_evaluations": 130,
  "incomplete": [],
  "public": true,
  "repo_id": "RL-Forgetting-Experiments-3/mbpp-code-sft-artifacts",
  "schema_version": 1,
  "status": "complete",
  "validated_evaluations": 130
}