RaindropA practical guide to choosing decision modelsGitHub ↗

Invented names · Use familiar names 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

Speaker_Alice: this bed is cushy as clouds Speaker_Robert: The bed is very uncomfortable

Question

Choose the interpretation or answer requested by the text.

Answer choices

  • A. Agrees
  • B. Disagrees

Reference answer: B.

Exact evaluated input as JSON
{
  "state": "Speaker_Alice: this bed is cushy as clouds\nSpeaker_Robert: The bed is very uncomfortable",
  "question": "Choose the interpretation or answer requested by the text.",
  "choices": {
    "A": "Agrees",
    "B": "Disagrees"
  }
}

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 B — Disagrees B — Disagrees Same answer
Decider v1.1 B — Disagrees B — Disagrees Same answer
Jev B — Disagrees B — Disagrees Same answer
Microsoft Decision-1 B — Disagrees B — Disagrees Same answer
Clef B — Disagrees B — Disagrees Same answer
Luna Decisions B — Disagrees B — Disagrees Same answer
Clef Flash B — Disagrees B — Disagrees Same answer
Strands Decider B — Disagrees B — Disagrees Same answer
GLiNER2.5 Decide B — Disagrees B — Disagrees Same answer
Laya B — Disagrees B — Disagrees Same answer
CalDec v1 B — Disagrees B — Disagrees Same answer
Identifiers and recorded conditions

Variant ID: pub/eec9dc6166eaca60cf90/unfamiliar_names/familiar

Input fingerprint: 3be56a5153d911a7a8b08010058db1f61be48aa5d1c497815b44766d08ac170c

Change Recorded version
Invented names Use familiar names

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

Type to search questions, variants and documentation.