Why can't dummy text be recognized?

Say to which category the next text is related - dummy text or real text:
Aenean sit amet ex tellus. Ut ac congue turpis. Suspendisse at sapien a lectus accumsan porttitor. Mauris varius justo volutpat, fermentum lorem non, ultricies est. Sed ullamcorper laoreet neque, in fringilla eros vehicula non.

ChatGPT
The text you provided is a real or “sense” text. It contains meaningful content that is understandable and conveys information.

This is a task that GPT-4 can do, but gpt-3.5-turbo fails at.

prompt to chatbot:

Discover and print the original language of the input passage and the confidence you have that you’ve identified the input language correctly, and report irregularities. You should attempt to discover if the input appears to be broken language placeholder text (in lorem ipsum style). Finally, translate the input text to high-quality English language. From the quality and success of the translation repeat analysis and detection of placeholder text and confidence.

input:
Aenean sit amet ex tellus. Ut ac congue turpis. Suspendisse at sapien a lectus accumsan porttitor. Mauris varius justo volutpat, fermentum lorem non, ultricies est. Sed ullamcorper laoreet neque, in fringilla eros vehicula non.

output:

Original Language: Latin
Confidence: 95%
Irregularities: The input text appears to be broken language placeholder text (in lorem ipsum style).

Translation to English:
Aenean sits with the former tellus. And with the congue turpis. Suspendisse at sapien a lectus accumsan porttitor. Mauris varius justo volutpat, fermentum lorem non, ultricies est. Sed ullamcorper laoreet neque, in fringilla eros vehicula non.

Quality and Success of Translation: The translation is of low quality and not successful, as the input text appears to be placeholder text (lorem ipsum style) and does not convey meaningful information.

1 Like

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.