here is an image generation prompt: wizard riding mountain bike on trail what questions…
- User
here is an image generation prompt:
wizard riding mountain bike on trail
what questions would you ask to make the prompt more detailed?
please include the question as well as 5 possible answers in the form of 1-2 keywords for each question. output your reply in a JSONL format:
[{"question": 'question', "answer_choices": ['answer1', 'answer2', 'answer3', 'answer4', 'answer5']}]
output 10 possible questions, for the answer choices try to make them varied and interesting.
- Assistant
Here are 10 possible questions with varied and interesting answer choices for the prompt "wizard riding mountain bike on trail":
[ { "question": "wizard appearance", "answer_choices": ["old", "young", "male", "female", "nonbinary"] }, { "question": "wizard outfit", "answer_choices": ["robes", "modern clothes", "armor", "hoodie", "cape"] }, { "question": "bike type", "answer_choices": ["mountain", "road", "BMX", "tandem", "unicycle"] }, { "question": "bike color", "answer_choices": ["red", "blue", "green", "rainbow", "camouflage"] }, { "question": "trail terrain", "answer_choices": ["dirt", "gravel", "snow", "ice", "sand"] }, { "question": "trail scenery", "answer_choices": ["forest", "desert", "jungle", "tundra", "meadow"] }, { "question": "wizard expression", "answer_choices": ["happy", "scared", "angry", "focused", "laughing"] }, { "question": "other characters", "answer_choices": ["dragon", "unicorn", "fairies", "troll", "gnome"] }, { "question": "time of day", "answer_choices": ["morning", "afternoon", "sunset", "night", "dawn"] }, { "question": "weather", "answer_choices": ["sunny", "rainy", "snowy", "windy", "foggy"] } ]
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.