Skip to content
TheChatArchiveThe Chat Archive
gpt-3.5-turbo-0301· OpenAIAi2 WildChat-4.8M (AI2)

Attention: Here is what makes a poll WELL FORMATTED: NO bad spelling and NO bad…

  1. User

    Attention: Here is what makes a poll WELL FORMATTED: NO bad spelling and NO bad punctuation NO bad grammar NO answers unrelated to question NO colloquial language NO unclear meaning This is a poll, NOT a quiz! NO multichoice answers NO duplicated answers

  2. Assistant

    Ensure that each question is clear and concise, and each answer option is mutually exclusive and collectively exhaustive. Use a balanced mix of positive and negative answer options, and include a "not sure" or "no opinion" option if appropriate. Avoid biased or leading language in your questions, and consider the order in which you present answer options to prevent any ordering effects. Finally, be transparent about the purpose of the poll and who will have access to the results.

  3. User

    What is the most surprising fact about Tim de Zarn? He did not pursue acting until he was 25 He went to a Catholic school in Ohio He lost his son in a car accident in 2007 He is married to Janine deZan Is the poll WELL FORMATTED?

  4. Assistant

    No, the poll is not well formatted. Here are the corrections:

    Question: What is the most surprising fact about Tim de Zarn? Answer options:

    • He did not pursue acting until he was 25
    • He went to a Catholic school in Ohio
    • He lost his son in a car accident in 2007
    • He is married to Janine deZan
    • Not sure

    Explanation: The question is clear and concise, and each answer option is mutually exclusive and collectively exhaustive. There is a balanced mix of positive and negative answer options, and a "not sure" option is included. However, it would be better to capitalize the first letter of "Janine deZan" for consistency and clarity. Additionally, the answer options could be arranged in a random order to prevent any ordering effects. Finally, it would be appropriate to disclose the purpose of the poll and who will have access to the results.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.