RaindropA practical guide to choosing decision modelsGitHub ↗

Learning and induction · question 27

Back to Learning and induction · All capabilities

Question ID: list_functions/631359375ff7f24f7e9b
Source: List Functions
Input format: adapted choices

Explore this question’s altered versions

Context

[{"input": "[39, 61, 49, 63, 89]", "output": "[39, 61, 49, 63, 61, 49, 63, 89]"}, {"input": "[5, 0, 67]", "output": "[5, 0, 0, 67]"}, {"input": "[25, 48, 15, 77]", "output": "[25, 48, 15, 48, 15, 77]"}, {"input": "[59, 71, 9]", "output": "[59, 71, 71, 9]"}, {"input": "[93, 11, 3, 6, 4]", "output": "[93, 11, 3, 6, 11, 3, 6, 4]"}, {"input": "[6, 8, 5]", "output": "[6, 8, 8, 5]"}, {"input": "[47, 45, 73]", "output": "[47, 45, 45, 73]"}, {"input": "[9, 74, 84, 1, 51]", "output": "[9, 74, 84, 1, 74, 84, 1, 51]"}]

Question

Apply the same inferred rule to this new input: [89, 0, 2, 97, 5]

Answer choices

  • A: [89, 0, 2, 97, 0, 2, 97, 5]
  • B: [89, 0, 2, 97, 5, 89, 0, 2]
  • C: [1, 2, 3, 4, 5, 1, 2, 3]
  • D: [97, 89, 5, 2, 0, 97, 89, 5]

Reference answer: [89, 0, 2, 97, 0, 2, 97, 5] (A)

Exact evaluated input as JSON
{
  "state": "[{\"input\": \"[39, 61, 49, 63, 89]\", \"output\": \"[39, 61, 49, 63, 61, 49, 63, 89]\"}, {\"input\": \"[5, 0, 67]\", \"output\": \"[5, 0, 0, 67]\"}, {\"input\": \"[25, 48, 15, 77]\", \"output\": \"[25, 48, 15, 48, 15, 77]\"}, {\"input\": \"[59, 71, 9]\", \"output\": \"[59, 71, 71, 9]\"}, {\"input\": \"[93, 11, 3, 6, 4]\", \"output\": \"[93, 11, 3, 6, 11, 3, 6, 4]\"}, {\"input\": \"[6, 8, 5]\", \"output\": \"[6, 8, 8, 5]\"}, {\"input\": \"[47, 45, 73]\", \"output\": \"[47, 45, 45, 73]\"}, {\"input\": \"[9, 74, 84, 1, 51]\", \"output\": \"[9, 74, 84, 1, 74, 84, 1, 51]\"}]",
  "question": "Apply the same inferred rule to this new input: [89, 0, 2, 97, 5]",
  "choices": {
    "A": "[89, 0, 2, 97, 0, 2, 97, 5]",
    "B": "[89, 0, 2, 97, 5, 89, 0, 2]",
    "C": "[1, 2, 3, 4, 5, 1, 2, 3]",
    "D": "[97, 89, 5, 2, 0, 97, 89, 5]"
  }
}

Recorded model answers

Model Answer Outcome
Astra [89, 0, 2, 97, 0, 2, 97, 5] Correct
Opus 5.5 [89, 0, 2, 97, 0, 2, 97, 5] Correct
GLiDE v1.2 [89, 0, 2, 97, 0, 2, 97, 5] Correct
Decider v1.1 [89, 0, 2, 97, 0, 2, 97, 5] Correct
Jev [89, 0, 2, 97, 0, 2, 97, 5] Correct
Microsoft Decision-1 [89, 0, 2, 97, 0, 2, 97, 5] Correct
Clef [89, 0, 2, 97, 0, 2, 97, 5] Correct
Luna Decisions [89, 0, 2, 97, 0, 2, 97, 5] Correct
Clef Flash [89, 0, 2, 97, 0, 2, 97, 5] Correct
Strands Decider [89, 0, 2, 97, 0, 2, 97, 5] Correct
GLiNER2.5 Decide [89, 0, 2, 97, 0, 2, 97, 5] Correct
Laya [89, 0, 2, 97, 0, 2, 97, 5] Correct
CalDec v1 [89, 0, 2, 97, 0, 2, 97, 5] Correct

Source and sharing

Licensed upstream task: c105. 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": "104/6",
    "kind": "source_file_record"
  },
  "source_revision": "0eae21f3d3014b92be34c80b23c9120a7b15a51f",
  "input_hash": "5ea21946603d54735fbac4785e61f0af233116c66a938200e5dd1f727d85f062"
}

Type to search questions, variants and documentation.