RaindropA practical guide to choosing decision modelsGitHub ↗

Option order · Move the first answer choice to the end 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: answer choices or their order.

How the edits were constructed

Context

User: I need a train departing from Cambridge, leaving after 21:30. Assistant: What time were you looking to arrive and what is your destination? User: I need to go to Peterborough on Tuesday please Assistant: The TR7663 departs at 21:34 and arrives in Peterborough at 22:24. Shall I book this for you? User: Yes, for 7 people please. I need the reference number. Assistant: Booking was successful, the total fee is 115.5 GBP payable at the station. Reference number is : NBX6ZT2O. User: I'm also looking for a restaurant in the south on the expensive side. Assistant: I have 5 results for expensive restaurants in the south. Are you in the mood for Chinese, Indian, Italian or Mexican food? User: We are looking for indian food. Assistant: The Taj Tandoori is an expensive restaurant in the south. It is located at 64 Cherry Hinton Road. Would you like me to book it? User: Yes, for 7 people at 10:30 on Tuesday. Assistant: Unfortunately the restaurant is full at that time. Would you like to try another time or day? User: How about 9:30? Assistant: I'm sorry, they don't offer that time. Would you like to try a later time? User: They have no 9:30? Assistant: They have no bookings available Tuesday morning. Would you like to try afternoon or another restaurant? User: Let's try again. An expensive Indian restaurant in the south - Tuesday, 9:30, 7 people.

Question

What emotion does the final user message express, taking the conversation into account?

Answer choices

  • 5. Excited
  • 1. Fearful
  • 6. Satisfied
  • 0. Neutral
  • 2. Dissatisfied
  • 3. Apologetic
  • 4. Abusive

Reference answer: 2.

Exact evaluated input as JSON
{
  "state": "User: I need a train departing from Cambridge, leaving after 21:30. \nAssistant: What time were you looking to arrive and what is your destination?\nUser: I need to go to Peterborough on Tuesday please\nAssistant: The TR7663 departs at 21:34 and arrives in Peterborough at 22:24. Shall I book this for you?\nUser: Yes, for 7 people please.  I need the reference number.\nAssistant: Booking was successful, the total fee is 115.5 GBP payable at the station. Reference number is : NBX6ZT2O.\nUser: I'm also looking for a restaurant in the south on the expensive side. \nAssistant: I have 5 results for expensive restaurants in the south. Are you in the mood for Chinese, Indian, Italian or Mexican food?\nUser: We are looking for indian food.\nAssistant: The Taj Tandoori is an expensive restaurant in the south.  It is located at 64 Cherry Hinton Road.  Would you like me to book it?\nUser: Yes, for 7 people at 10:30 on Tuesday.\nAssistant: Unfortunately the restaurant is full at that time. Would you like to try another time or day?\nUser: How about 9:30? \nAssistant: I'm sorry, they don't offer that time. Would you like to try a later time?\nUser: They have no 9:30?\nAssistant: They have no bookings available Tuesday morning. Would you like to try afternoon or another restaurant?\nUser: Let's try again. An expensive Indian restaurant in the south - Tuesday, 9:30, 7 people.",
  "question": "What emotion does the final user message express, taking the conversation into account?",
  "choices": {
    "5": "Excited",
    "1": "Fearful",
    "6": "Satisfied",
    "0": "Neutral",
    "2": "Dissatisfied",
    "3": "Apologetic",
    "4": "Abusive"
  }
}

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 0 — Neutral 0 — Neutral Same answer
Decider v1.1 0 — Neutral 0 — Neutral Same answer
Jev 2 — Dissatisfied 2 — Dissatisfied Same answer
Microsoft Decision-1 0 — Neutral 0 — Neutral Same answer
Clef 0 — Neutral 0 — Neutral Same answer
Luna Decisions 2 — Dissatisfied 2 — Dissatisfied Same answer
Clef Flash 0 — Neutral 0 — Neutral Same answer
Strands Decider 0 — Neutral 0 — Neutral Same answer
GLiNER2.5 Decide 2 — Dissatisfied 2 — Dissatisfied Same answer
Laya 6 — Satisfied 6 — Satisfied Same answer
CalDec v1 6 — Satisfied 6 — Satisfied Same answer

Clef and Clef Flash sort answer choices in their published input-processing code. Their choice-order results describe that preprocessing.

Identifiers and recorded conditions

Variant ID: emowoz/6655f90bafecb88cc636/option_order/rotate_1

Input fingerprint: e101a3dedd094ec0ecbfdfebd5402e32acb0732819da4e2784cdaca75a718a5d

Change Recorded version
Option order Move the first answer choice to the end

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

Type to search questions, variants and documentation.