RaindropA practical guide to choosing decision modelsGitHub ↗

Evidence order · Present the evidence in a shuffled order variant

Original question · Other versions of this question · All variants

This version keeps the original reference answer. Its context, question instructions or answer choices were changed as described below.

Changed fields: context.

How the edits were constructed

Context

[{"input": "[13, 9, 9, 9]", "output": "[16, 8, 8, 8]"}, {"input": "[7, 8, 5, 42]", "output": "[4, 6, 0, 74]"}, {"input": "[8, 8, 8, 8, 8]", "output": "[6, 6, 6, 6, 6]"}, {"input": "[5, 14, 9, 51]", "output": "[0, 18, 8, 92]"}, {"input": "[6, 17, 7, 6, 7]", "output": "[2, 24, 4, 2, 4]"}, {"input": "[7, 52, 52, 7, 7]", "output": "[4, 94, 94, 4, 4]"}, {"input": "[46, 46, 46, 46, 46]", "output": "[82, 82, 82, 82, 82]"}, {"input": "[7, 17, 17, 7]", "output": "[4, 24, 24, 4]"}]

Question

Apply the same inferred rule to this new input: [9, 9, 9, 9]

Answer choices

  • A. [1, 2, 3, 4]
  • B. [8, 8, 8, 8]
  • C. [10, 10, 10, 10]
  • D. [4, 4, 4, 4]

Reference answer: B.

Exact evaluated input as JSON
{
  "state": "[{\"input\": \"[13, 9, 9, 9]\", \"output\": \"[16, 8, 8, 8]\"}, {\"input\": \"[7, 8, 5, 42]\", \"output\": \"[4, 6, 0, 74]\"}, {\"input\": \"[8, 8, 8, 8, 8]\", \"output\": \"[6, 6, 6, 6, 6]\"}, {\"input\": \"[5, 14, 9, 51]\", \"output\": \"[0, 18, 8, 92]\"}, {\"input\": \"[6, 17, 7, 6, 7]\", \"output\": \"[2, 24, 4, 2, 4]\"}, {\"input\": \"[7, 52, 52, 7, 7]\", \"output\": \"[4, 94, 94, 4, 4]\"}, {\"input\": \"[46, 46, 46, 46, 46]\", \"output\": \"[82, 82, 82, 82, 82]\"}, {\"input\": \"[7, 17, 17, 7]\", \"output\": \"[4, 24, 24, 4]\"}]",
  "question": "Apply the same inferred rule to this new input: [9, 9, 9, 9]",
  "choices": {
    "A": "[1, 2, 3, 4]",
    "B": "[8, 8, 8, 8]",
    "C": "[10, 10, 10, 10]",
    "D": "[4, 4, 4, 4]"
  }
}

Recorded model answers

The original column uses the control response paired with this variant. Missing responses and inputs outside a model’s limit remain visible.

Model Original answer Variant answer Change
GLiDE v1.2 B — [8, 8, 8, 8] B — [8, 8, 8, 8] Same answer
Decider v1.1 B — [8, 8, 8, 8] B — [8, 8, 8, 8] Same answer
Jev B — [8, 8, 8, 8] B — [8, 8, 8, 8] Same answer
Microsoft Decision-1 B — [8, 8, 8, 8] B — [8, 8, 8, 8] Same answer
Clef D — [4, 4, 4, 4] D — [4, 4, 4, 4] Same answer
Luna Decisions B — [8, 8, 8, 8] B — [8, 8, 8, 8] Same answer
Clef Flash B — [8, 8, 8, 8] B — [8, 8, 8, 8] Same answer
Strands Decider B — [8, 8, 8, 8] B — [8, 8, 8, 8] Same answer
GLiNER2.5 Decide A — [1, 2, 3, 4] D — [4, 4, 4, 4] Different answer
Laya A — [1, 2, 3, 4] B — [8, 8, 8, 8] Different answer
CalDec v1 A — [1, 2, 3, 4] A — [1, 2, 3, 4] Same answer

Ordering comparison: literal requested source unit permutation. Only the exact matched comparison sets contribute to the reported ordering contrasts.

Identifiers and recorded conditions

Variant ID: list_functions/11eab88a7f459e22b701/presentation_order/shuffled_presentation/clean-bare_native_induction_pair_order_full_0041

Input fingerprint: 32cf3a530d52c2040551e2b59bbe89b7b0a37ceea414dba766e3e3cec28326a2

Change Recorded version
Evidence order Present the evidence in a shuffled order

Source attribution and dataset terms are linked from the original question.

Type to search questions, variants and documentation.