Reasoning-Fce-Greedy-Allocator

The model receives the prompt (and optional system message). The run uses scorer rubric_json_metrics with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.

Return JSON only with keys answer, evidence, constraints. A greedy allocator always chooses the largest task first. Find a small capacity/task set where this fails to maximize filled capacity.

{
  "metrics": {
    "accuracy": {
      "checks": [
        {
          "contains": [
            "capacity"
          ]
        },
        {
          "contains": [
            "largest task"
          ]
        },
        {
          "contains": [
            "counterexample"
          ]
        }
      ]
    },
    "evidence": {
      "checks": [
        {
          "contains": [
            "greedy"
          ]
        },
        {
          "contains": [
            "maximize"
          ]
        },
        {
          "contains": [
            "filled"
          ]
        }
      ]
    },
    "constraint": {
      "checks": [
        {
          "contains": [
            "smaller combination"
          ]
        },
        {
          "contains": [
            "knapsack"
          ]
        },
        {
          "contains": [
            "fails"
          ]
        }
      ]
    }
  }
}