RaindropA practical guide to choosing decision modelsGitHub ↗

Numerical cognition · question 62

Back to Numerical cognition · All capabilities

Question ID: bbh/40eceedeab129fcf9d44
Source: BIG-Bench Hard selected tasks
Input format: adapted choices

Explore this question’s altered versions

Question

((3 + 3 - -4 - 2) + (-2 - -7 * -2 * 1)) =

Answer choices

  • A: 8
  • B: 14
  • C: -1
  • D: -8

Reference answer: -8 (D)

Exact evaluated input as JSON
{
  "state": "",
  "question": "((3 + 3 - -4 - 2) + (-2 - -7 * -2 * 1)) =",
  "choices": {
    "A": "8",
    "B": "14",
    "C": "-1",
    "D": "-8"
  }
}

Recorded model answers

Model Answer Outcome
Astra -8 Correct
Opus 5.5 -8 Correct
GLiDE v1.2 -8 Correct
Decider v1.1 -8 Correct
Jev -8 Correct
Microsoft Decision-1 -8 Correct
Clef -1 Incorrect
Luna Decisions -1 Incorrect
Clef Flash -1 Incorrect
Strands Decider 8 Incorrect
GLiNER2.5 Decide 8 Incorrect
Laya -8 Correct
CalDec v1 -1 Incorrect

Source and sharing

Attribution: Mirac Suzgun and BIG-Bench Hard authors; BIG-bench contributors.
Terms: MIT AND 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": "bbh/multistep_arithmetic_two.json",
    "record": "121",
    "kind": "source_file_record"
  },
  "source_revision": "9ee07bd481feebf959a6b59d61ea57bdcf30964d",
  "input_hash": "36f8f7faafb259738d8fd2ae60af7fdee4c67d4d650b29b32c8e511266ce95c6"
}

Type to search questions, variants and documentation.