{
  "models": [
    {
      "tag": "baseline",
      "seconds": 105.29783153533936,
      "within_template": {
        "n": 36,
        "correct": 0,
        "unparsed": 23,
        "hit_limit": 17,
        "mean_generated_tokens": 103.5
      },
      "probe": {
        "n": 12,
        "correct": 10,
        "unparsed": 2,
        "hit_limit": 2,
        "mean_generated_tokens": 35.5
      },
      "tasks": {
        "discount": {
          "n": 12,
          "correct": 0
        },
        "inventory": {
          "n": 12,
          "correct": 0
        },
        "percentage": {
          "n": 12,
          "correct": 0
        },
        "probe": {
          "n": 12,
          "correct": 10
        }
      }
    },
    {
      "tag": "explained-seed17",
      "seconds": 85.88354325294495,
      "within_template": {
        "n": 36,
        "correct": 32,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 46.97222222222222
      },
      "probe": {
        "n": 12,
        "correct": 11,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 14.666666666666666
      },
      "tasks": {
        "discount": {
          "n": 12,
          "correct": 11
        },
        "inventory": {
          "n": 12,
          "correct": 12
        },
        "percentage": {
          "n": 12,
          "correct": 9
        },
        "probe": {
          "n": 12,
          "correct": 11
        }
      },
      "training": {
        "steps": 72,
        "supervised_tokens": 13974,
        "seconds": 185.7444965839386,
        "trainable_parameters": 540672,
        "adapter_l1_change": 1710.9576642513275
      },
      "versus_baseline": {
        "improved": [
          "discount-170-15-180",
          "discount-200-30-200",
          "discount-230-10-220",
          "discount-180-25-180",
          "discount-170-35-180",
          "discount-150-30-180",
          "discount-150-25-180",
          "discount-230-20-200",
          "discount-220-10-180",
          "discount-230-15-220",
          "discount-230-25-180",
          "inventory-8-6-8",
          "inventory-8-11-2",
          "inventory-9-7-7",
          "inventory-3-12-5",
          "inventory-8-9-5",
          "inventory-9-4-8",
          "inventory-9-11-8",
          "inventory-4-7-4",
          "inventory-9-12-3",
          "inventory-12-7-3",
          "inventory-3-10-3",
          "inventory-7-13-7",
          "percentage-800-50",
          "percentage-900-10",
          "percentage-700-10",
          "percentage-200-20",
          "percentage-1000-30",
          "percentage-600-30",
          "percentage-1000-10",
          "percentage-400-10",
          "percentage-400-50",
          "probe-08",
          "probe-11"
        ],
        "regressed": [
          "probe-05"
        ]
      }
    },
    {
      "tag": "answer_only-seed17",
      "seconds": 19.335241556167603,
      "within_template": {
        "n": 36,
        "correct": 16,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 7.027777777777778
      },
      "probe": {
        "n": 12,
        "correct": 11,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 6.666666666666667
      },
      "tasks": {
        "discount": {
          "n": 12,
          "correct": 5
        },
        "inventory": {
          "n": 12,
          "correct": 5
        },
        "percentage": {
          "n": 12,
          "correct": 6
        },
        "probe": {
          "n": 12,
          "correct": 11
        }
      },
      "training": {
        "steps": 72,
        "supervised_tokens": 2349,
        "seconds": 128.9539499282837,
        "trainable_parameters": 540672,
        "adapter_l1_change": 1103.5804759263992
      },
      "versus_baseline": {
        "improved": [
          "discount-170-15-180",
          "discount-200-20-220",
          "discount-170-35-180",
          "discount-150-30-180",
          "discount-150-25-180",
          "inventory-8-11-2",
          "inventory-4-7-4",
          "inventory-9-12-3",
          "inventory-12-7-3",
          "inventory-3-10-3",
          "percentage-800-50",
          "percentage-700-10",
          "percentage-500-10",
          "percentage-1000-10",
          "percentage-400-10",
          "percentage-600-10",
          "probe-08",
          "probe-11"
        ],
        "regressed": [
          "probe-04"
        ]
      }
    },
    {
      "tag": "explained-seed29",
      "seconds": 89.09294891357422,
      "within_template": {
        "n": 36,
        "correct": 35,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 47.47222222222222
      },
      "probe": {
        "n": 12,
        "correct": 11,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 16.5
      },
      "tasks": {
        "discount": {
          "n": 12,
          "correct": 12
        },
        "inventory": {
          "n": 12,
          "correct": 12
        },
        "percentage": {
          "n": 12,
          "correct": 11
        },
        "probe": {
          "n": 12,
          "correct": 11
        }
      },
      "training": {
        "steps": 72,
        "supervised_tokens": 13974,
        "seconds": 203.2175588607788,
        "trainable_parameters": 540672,
        "adapter_l1_change": 1696.6415810585022
      },
      "versus_baseline": {
        "improved": [
          "discount-170-15-180",
          "discount-200-30-200",
          "discount-200-20-220",
          "discount-230-10-220",
          "discount-180-25-180",
          "discount-170-35-180",
          "discount-150-30-180",
          "discount-150-25-180",
          "discount-230-20-200",
          "discount-220-10-180",
          "discount-230-15-220",
          "discount-230-25-180",
          "inventory-8-6-8",
          "inventory-8-11-2",
          "inventory-9-7-7",
          "inventory-3-12-5",
          "inventory-8-9-5",
          "inventory-9-4-8",
          "inventory-9-11-8",
          "inventory-4-7-4",
          "inventory-9-12-3",
          "inventory-12-7-3",
          "inventory-3-10-3",
          "inventory-7-13-7",
          "percentage-800-50",
          "percentage-900-10",
          "percentage-700-10",
          "percentage-200-20",
          "percentage-500-10",
          "percentage-1000-30",
          "percentage-600-30",
          "percentage-1000-10",
          "percentage-400-10",
          "percentage-600-10",
          "percentage-400-50",
          "probe-08",
          "probe-11"
        ],
        "regressed": [
          "probe-05"
        ]
      }
    },
    {
      "tag": "answer_only-seed29",
      "seconds": 20.039268016815186,
      "within_template": {
        "n": 36,
        "correct": 17,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 7.361111111111111
      },
      "probe": {
        "n": 12,
        "correct": 10,
        "unparsed": 0,
        "hit_limit": 0,
        "mean_generated_tokens": 6.666666666666667
      },
      "tasks": {
        "discount": {
          "n": 12,
          "correct": 7
        },
        "inventory": {
          "n": 12,
          "correct": 6
        },
        "percentage": {
          "n": 12,
          "correct": 4
        },
        "probe": {
          "n": 12,
          "correct": 10
        }
      },
      "training": {
        "steps": 72,
        "supervised_tokens": 2349,
        "seconds": 138.94542026519775,
        "trainable_parameters": 540672,
        "adapter_l1_change": 1074.1055912971497
      },
      "versus_baseline": {
        "improved": [
          "discount-200-30-200",
          "discount-230-10-220",
          "discount-180-25-180",
          "discount-230-20-200",
          "discount-220-10-180",
          "discount-230-15-220",
          "discount-230-25-180",
          "inventory-8-6-8",
          "inventory-9-7-7",
          "inventory-8-9-5",
          "inventory-9-12-3",
          "inventory-12-7-3",
          "inventory-3-10-3",
          "percentage-700-10",
          "percentage-500-10",
          "percentage-1000-10",
          "percentage-600-10",
          "probe-08",
          "probe-11"
        ],
        "regressed": [
          "probe-04",
          "probe-05"
        ]
      }
    }
  ],
  "paired_disagreements": [
    {
      "seed": 17,
      "explained_only_correct": [
        "discount-200-30-200",
        "discount-230-10-220",
        "discount-180-25-180",
        "discount-230-20-200",
        "discount-220-10-180",
        "discount-230-15-220",
        "discount-230-25-180",
        "inventory-8-6-8",
        "inventory-9-7-7",
        "inventory-3-12-5",
        "inventory-8-9-5",
        "inventory-9-4-8",
        "inventory-9-11-8",
        "inventory-7-13-7",
        "percentage-900-10",
        "percentage-200-20",
        "percentage-1000-30",
        "percentage-600-30",
        "percentage-400-50",
        "probe-04"
      ],
      "answer_only_correct": [
        "discount-200-20-220",
        "percentage-500-10",
        "percentage-600-10",
        "probe-05"
      ]
    },
    {
      "seed": 29,
      "explained_only_correct": [
        "discount-170-15-180",
        "discount-200-20-220",
        "discount-170-35-180",
        "discount-150-30-180",
        "discount-150-25-180",
        "inventory-8-11-2",
        "inventory-3-12-5",
        "inventory-9-4-8",
        "inventory-9-11-8",
        "inventory-4-7-4",
        "inventory-7-13-7",
        "percentage-800-50",
        "percentage-900-10",
        "percentage-200-20",
        "percentage-1000-30",
        "percentage-600-30",
        "percentage-400-10",
        "percentage-400-50",
        "probe-04"
      ],
      "answer_only_correct": []
    }
  ]
}
