Skip to content
TheChatArchiveThe Chat Archive
Claudeclaude-instant-1· AnthropicHugging Face Arena Human Preference 55K

point out the typora. \caption{ Natural Language Generation Results from Prompts in NLG…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with gpt-4-0314’s answer.
  1. User

    point out the typora.

      \caption{ Natural Language Generation Results from Prompts in NLG tasks (translated from Chinese to English). The models generate the masked spans inside the prompts. We replace "\textless{}{[}MASK{]}\textgreater{}" in prompts with generated span. The result of "GUR-LM" are ignored for bad and long sentences. All models pretrained with LM task have a same and fluent generated result. We write Artificial Prompts and mask random optional words into Auto Prompts. Comparing to the result from the Artificial Prompts, the sentences from the Auto Prompts have more diversity. }      \label{tab:Prompts}
    
  2. Claude Assistant

    Here are the typos I found in the given text:

    1. typora should be typo.
    2. "\textless{}{[}MASK{]}\textgreater{}" should be "\textless{}[MASK]\textgreater{}"
    3. "GUR-LM" should be "GTP-LM"
    4. "All models pretrained with LM task have a same and fluent generated result." should be "All models pretrained with the LM task have similar and fluent generated results."
    5. "Comparing to the result from the Artificial Prompts," should be "Compared to the results from the Artificial Prompts,"

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