{
  "run": "smoke_inpolicy_vm_echo",
  "args": {
    "run_name": "smoke_inpolicy_vm_echo",
    "model_name": "Qwen/Qwen3-4B",
    "seed": 91,
    "device": "",
    "seed_train_size": 8,
    "unlabeled_train_size": 12,
    "full_supervised_size": 12,
    "val_size": 8,
    "fresh_size": 8,
    "hard_size": 8,
    "max_arith_steps": 4,
    "qwen_batch_size": 2,
    "d_model": 64,
    "compiler_layers": 1,
    "echo_layers": 1,
    "heads": 4,
    "dropout": 0.1,
    "batch_size": 4,
    "group_batch_size": 2,
    "lr": 0.0005,
    "distill_lr": 0.00025,
    "echo_lr": 0.00025,
    "answer_weight": 0.2,
    "seed_epochs": 1,
    "distill_epochs": 1,
    "echo_epochs": 1,
    "gold_epochs": 1,
    "full_supervised_epochs": 1,
    "inpolicy_rounds": 1,
    "search_topk": 2,
    "max_two_arg_pairs": 2,
    "max_candidates": 32,
    "echo_loss_weight": 0.1,
    "correct_loss_weight": 1.0,
    "group_loss_weight": 1.0,
    "valid_loss_weight": 0.2,
    "final_loss_weight": 0.2,
    "trace_top_loss_weight": 0.1,
    "trace_depth_loss_weight": 0.05
  },
  "results": [
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "seed_supervised",
      "split": "val_mixed",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "seed_supervised",
      "split": "fresh_standard",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "seed_supervised",
      "split": "fresh_paraphrase",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "seed_supervised",
      "split": "fresh_paired",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "seed_supervised",
      "split": "hard_composition",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "answer_verified_distill",
      "split": "val_mixed",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "answer_verified_distill",
      "split": "fresh_standard",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "answer_verified_distill",
      "split": "fresh_paraphrase",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "answer_verified_distill",
      "split": "fresh_paired",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "answer_verified_distill",
      "split": "hard_composition",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "inpolicy_vm_echo_distill",
      "split": "val_mixed",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": 0.0,
      "echo_rerank_valid_rate": 0.0,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": 0.0,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": 1.0,
      "echo_valid_pred_accuracy": 0.33653846153846156,
      "echo_final_pred_accuracy": 0.0,
      "echo_trace_top_accuracy": 0.0,
      "echo_trace_depth_accuracy": 0.05
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "inpolicy_vm_echo_distill",
      "split": "fresh_standard",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": 0.0,
      "echo_rerank_valid_rate": 0.0,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": 0.0,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": 1.0,
      "echo_valid_pred_accuracy": 0.36778846153846156,
      "echo_final_pred_accuracy": 0.0,
      "echo_trace_top_accuracy": 0.0,
      "echo_trace_depth_accuracy": 0.05
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "inpolicy_vm_echo_distill",
      "split": "fresh_paraphrase",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": 0.0,
      "echo_rerank_valid_rate": 0.0,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": 0.0,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": 1.0,
      "echo_valid_pred_accuracy": 0.5384615384615384,
      "echo_final_pred_accuracy": 0.0,
      "echo_trace_top_accuracy": 0.0,
      "echo_trace_depth_accuracy": 0.05
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "inpolicy_vm_echo_distill",
      "split": "fresh_paired",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": 0.0,
      "echo_rerank_valid_rate": 0.0,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": 0.0,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": 1.0,
      "echo_valid_pred_accuracy": 0.9567307692307693,
      "echo_final_pred_accuracy": 0.0,
      "echo_trace_top_accuracy": 0.0,
      "echo_trace_depth_accuracy": 0.05
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "inpolicy_vm_echo_distill",
      "split": "hard_composition",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": 0.0,
      "echo_rerank_valid_rate": 0.0,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": 0.0,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": 1.0,
      "echo_valid_pred_accuracy": 0.5865384615384616,
      "echo_final_pred_accuracy": 0.0,
      "echo_trace_top_accuracy": 0.0,
      "echo_trace_depth_accuracy": 0.05
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "gold_trace_distill",
      "split": "val_mixed",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "gold_trace_distill",
      "split": "fresh_standard",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "gold_trace_distill",
      "split": "fresh_paraphrase",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "gold_trace_distill",
      "split": "fresh_paired",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "gold_trace_distill",
      "split": "hard_composition",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "full_supervised",
      "split": "val_mixed",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "full_supervised",
      "split": "fresh_standard",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "full_supervised",
      "split": "fresh_paraphrase",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "full_supervised",
      "split": "fresh_paired",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    },
    {
      "run": "smoke_inpolicy_vm_echo",
      "phase": "full_supervised",
      "split": "hard_composition",
      "n": 8,
      "direct_accuracy": 0.0,
      "answer_search_accuracy": 0.0,
      "oracle_accuracy": 0.0,
      "echo_rerank_accuracy": null,
      "echo_rerank_valid_rate": null,
      "direct_valid_rate": 0.0,
      "candidate_valid_rate": 0.0,
      "mean_candidates": 52.0,
      "found_rate": 0.0,
      "program_exact": 0.0,
      "echo_program_exact": null,
      "echo_gap_recovered": null,
      "echo_correct_pred_accuracy": null,
      "echo_valid_pred_accuracy": null,
      "echo_final_pred_accuracy": null,
      "echo_trace_top_accuracy": null,
      "echo_trace_depth_accuracy": null
    }
  ]
}