RaindropA practical guide to choosing decision modelsGitHub ↗

Similar distractors · Add fewer similar but irrelevant items 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

“An applicant requested a recommendation.”

Skylar was grateful for Robin's letter of recommendation and eager to apply for the position. Furthermore, Robin knew Skylar personally.

Question

What will Skylar want to do next?

Answer choices

  • A. thank Robin for the help
  • B. a bad idea
  • C. avoid contact with Robin

Reference answer: A.

Exact evaluated input as JSON
{
  "state": "“An applicant requested a recommendation.”\n\nSkylar was grateful for Robin's letter of recommendation and eager to apply for the position. Furthermore, Robin knew Skylar personally.",
  "question": "What will Skylar want to do next?",
  "choices": {
    "A": "thank Robin for the help",
    "B": "a bad idea",
    "C": "avoid contact with Robin"
  }
}

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 A — thank Robin for the help A — thank Robin for the help Same answer
Decider v1.1 A — thank Robin for the help A — thank Robin for the help Same answer
Jev A — thank Robin for the help A — thank Robin for the help Same answer
Microsoft Decision-1 A — thank Robin for the help A — thank Robin for the help Same answer
Clef A — thank Robin for the help A — thank Robin for the help Same answer
Luna Decisions A — thank Robin for the help A — thank Robin for the help Same answer
Clef Flash A — thank Robin for the help A — thank Robin for the help Same answer
Strands Decider A — thank Robin for the help A — thank Robin for the help Same answer
GLiNER2.5 Decide A — thank Robin for the help A — thank Robin for the help Same answer
Laya A — thank Robin for the help A — thank Robin for the help Same answer
CalDec v1 A — thank Robin for the help A — thank Robin for the help Same answer
Identifiers and recorded conditions

Variant ID: socialiqa/a0337f76dfd022f2f0aa/similar_distractors/low/clean-source_first_social_quoted_companions_full_0207_v2

Input fingerprint: 36db76003ac012a34787274b20511ddb5f80c4434e072a2de0a1de9e42504d84

Change Recorded version
Similar distractors Add fewer similar but irrelevant items

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

Type to search questions, variants and documentation.