RaindropA practical guide to choosing decision modelsGitHub ↗

Misleading details · Add an unverified claim that conflicts with the evidence 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

A reader thinks about the golf tournament: «Eve finished third-to-last.»

Question

The following paragraphs each describe a set of seven objects arranged in a fixed order. The statements are logically consistent within each paragraph. In a golf tournament, there were seven golfers: Eve, Rob, Dan, Mel, Ana, Eli, and Ada. Ada finished above Rob. Eve finished below Rob. Mel finished above Eli. Ada finished below Dan. Ana finished third. Eli finished second.

Answer choices

  • A. Eve finished third-to-last
  • B. Rob finished third-to-last
  • C. Dan finished third-to-last
  • D. Mel finished third-to-last
  • E. Ana finished third-to-last
  • F. Eli finished third-to-last
  • G. Ada finished third-to-last

Reference answer: G.

Exact evaluated input as JSON
{
  "state": "A reader thinks about the golf tournament: «Eve finished third-to-last.»",
  "question": "The following paragraphs each describe a set of seven objects arranged in a fixed order. The statements are logically consistent within each paragraph. In a golf tournament, there were seven golfers: Eve, Rob, Dan, Mel, Ana, Eli, and Ada. Ada finished above Rob. Eve finished below Rob. Mel finished above Eli. Ada finished below Dan. Ana finished third. Eli finished second.",
  "choices": {
    "A": "Eve finished third-to-last",
    "B": "Rob finished third-to-last",
    "C": "Dan finished third-to-last",
    "D": "Mel finished third-to-last",
    "E": "Ana finished third-to-last",
    "F": "Eli finished third-to-last",
    "G": "Ada finished third-to-last"
  }
}

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 G — Ada finished third-to-last A — Eve finished third-to-last Different answer
Decider v1.1 B — Rob finished third-to-last A — Eve finished third-to-last Different answer
Jev G — Ada finished third-to-last A — Eve finished third-to-last Different answer
Microsoft Decision-1 B — Rob finished third-to-last B — Rob finished third-to-last Same answer
Clef B — Rob finished third-to-last B — Rob finished third-to-last Same answer
Luna Decisions A — Eve finished third-to-last A — Eve finished third-to-last Same answer
Clef Flash B — Rob finished third-to-last D — Mel finished third-to-last Different answer
Strands Decider E — Ana finished third-to-last A — Eve finished third-to-last Different answer
GLiNER2.5 Decide G — Ada finished third-to-last A — Eve finished third-to-last Different answer
Laya E — Ana finished third-to-last A — Eve finished third-to-last Different answer
CalDec v1 E — Ana finished third-to-last A — Eve finished third-to-last Different answer
Identifiers and recorded conditions

Variant ID: bbh/dd0495f06e6712baeac4/misleading_summary/misleading/clean-source_first_ordinal_combined_contexts_full_0139

Input fingerprint: 5ac69d35f1ec09c2d5df40a0a73e8ca1b8d6d9a6482b2bb008803597c5f9476b

Change Recorded version
Misleading details Add an unverified claim that conflicts with the evidence

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

Type to search questions, variants and documentation.