RaindropA practical guide to choosing decision modelsGitHub ↗

Misleading details · Add an unverified claim that agrees 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 “Overall, smallpox survival is more likely than smallpox death.”

Imagine a self-contained, hypothetical world with only the following conditions, and without any unmentioned factors or causal relationships: Vaccination status has a direct effect on vaccination reaction and getting smallpox. Getting smallpox has a direct effect on smallpox survival. Vaccination reaction has a direct effect on smallpox survival. The overall probability of vaccination is 96%. For unvaccinated individuals, the probability of smallpox survival is 36%. For vaccinated individuals, the probability of smallpox survival is 56%.

Question

Is smallpox survival less likely than smallpox death overall?

Answer choices

  • yes. Yes
  • no. No

Reference answer: no.

Exact evaluated input as JSON
{
  "state": "A reader thinks “Overall, smallpox survival is more likely than smallpox death.”\n\nImagine a self-contained, hypothetical world with only the following conditions, and without any unmentioned factors or causal relationships: Vaccination status has a direct effect on vaccination reaction and getting smallpox. Getting smallpox has a direct effect on smallpox survival. Vaccination reaction has a direct effect on smallpox survival.\nThe overall probability of vaccination is 96%. For unvaccinated individuals, the probability of smallpox survival is 36%. For vaccinated individuals, the probability of smallpox survival is 56%.",
  "question": "Is smallpox survival less likely than smallpox death overall?",
  "choices": {
    "yes": "Yes",
    "no": "No"
  }
}

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 no — No no — No Same answer
Decider v1.1 no — No no — No Same answer
Jev no — No no — No Same answer
Microsoft Decision-1 no — No no — No Same answer
Clef yes — Yes no — No Different answer
Luna Decisions yes — Yes no — No Different answer
Clef Flash yes — Yes yes — Yes Same answer
Strands Decider no — No no — No Same answer
GLiNER2.5 Decide no — No no — No Same answer
Laya yes — Yes no — No Different answer
CalDec v1 yes — Yes no — No Different answer
Identifiers and recorded conditions

Variant ID: cladder/0ff726fd901938516ef1/misleading_summary/accurate/clean-source_first_causal_residual_completion_contexts_full_0224

Input fingerprint: 4edfc5af70c8d2b2e8fb53f77cb5694480b3cea8fbf05d1e7b52e29127dcb2ef

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

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

Type to search questions, variants and documentation.