RaindropA practical guide to choosing decision modelsGitHub ↗

Learning and induction · question 65

Back to Learning and induction · All capabilities

Question ID: list_functions/25cf4f190b25efb33fc0
Source: List Functions
Input format: adapted choices

Explore this question’s altered versions

Context

[{"input": "[2, 0, 5, 1, 3, 8, 9, 4, 7]", "output": "[2, 0]"}, {"input": "[3, 7, 9, 8, 6, 0, 4]", "output": "[3, 7]"}, {"input": "[1, 0, 8]", "output": "[1, 0]"}, {"input": "[0, 2, 7, 4, 5, 9, 3, 1, 8]", "output": "[0, 2]"}, {"input": "[8, 6, 4, 5, 7]", "output": "[8, 6]"}, {"input": "[2, 4, 7, 0, 8, 3, 5, 6, 1, 9]", "output": "[2, 4]"}, {"input": "[5]", "output": "[5]"}, {"input": "[6, 8, 3, 2, 9, 4, 5, 1]", "output": "[6, 8]"}]

Question

Apply the same inferred rule to this new input: [7, 1, 4]

Answer choices

  • A: [6, 0]
  • B: [4, 1]
  • C: [7, 1]
  • D: [4, 7]

Reference answer: [7, 1] (C)

Exact evaluated input as JSON
{
  "state": "[{\"input\": \"[2, 0, 5, 1, 3, 8, 9, 4, 7]\", \"output\": \"[2, 0]\"}, {\"input\": \"[3, 7, 9, 8, 6, 0, 4]\", \"output\": \"[3, 7]\"}, {\"input\": \"[1, 0, 8]\", \"output\": \"[1, 0]\"}, {\"input\": \"[0, 2, 7, 4, 5, 9, 3, 1, 8]\", \"output\": \"[0, 2]\"}, {\"input\": \"[8, 6, 4, 5, 7]\", \"output\": \"[8, 6]\"}, {\"input\": \"[2, 4, 7, 0, 8, 3, 5, 6, 1, 9]\", \"output\": \"[2, 4]\"}, {\"input\": \"[5]\", \"output\": \"[5]\"}, {\"input\": \"[6, 8, 3, 2, 9, 4, 5, 1]\", \"output\": \"[6, 8]\"}]",
  "question": "Apply the same inferred rule to this new input: [7, 1, 4]",
  "choices": {
    "A": "[6, 0]",
    "B": "[4, 1]",
    "C": "[7, 1]",
    "D": "[4, 7]"
  }
}

Recorded model answers

Model Answer Outcome
Astra [7, 1] Correct
Opus 5.5 [7, 1] Correct
GLiDE v1.2 [7, 1] Correct
Decider v1.1 [7, 1] Correct
Jev [7, 1] Correct
Microsoft Decision-1 [7, 1] Correct
Clef [7, 1] Correct
Luna Decisions [7, 1] Correct
Clef Flash [7, 1] Correct
Strands Decider [7, 1] Correct
GLiNER2.5 Decide [6, 0] Incorrect
Laya [4, 1] Incorrect
CalDec v1 [6, 0] Incorrect

Source and sharing

Licensed upstream task: c007. All eight supplied examples and the evaluated test input/output match the original BIG-bench data.

Attribution: Joshua S. Rule, Steven T. Piantadosi, Joshua B. Tenenbaum; BIG-bench contributors; Linlu Qiu et al. (selection).
Terms: Apache-2.0. Source sharing decision.
The capability assignment and choice-format normalization are part of this experiment. Source data retain their own terms.

Exact source locator and input fingerprint
{
  "source_locator": {
    "path": "data/list_function.jsonl",
    "record": "6/6",
    "kind": "source_file_record"
  },
  "source_revision": "0eae21f3d3014b92be34c80b23c9120a7b15a51f",
  "input_hash": "823864b4708c24b975d351f2289892f06dfb8308db218bdd9c7ac0e9e5f95f12"
}

Type to search questions, variants and documentation.