Why am I reading this in spanish instead of the original post language?

[nat (Natalie T)]

This is a little quirk with the LLM now, but should be fixed with better prompts soon.

Currently the prompt for determining a post’s locale is as such:

You will be given a piece of text, and your task is to detect the locale (language) of the text and return it in a specific JSON format.

To complete this task, follow these steps:

1. Carefully read and analyze the provided text.
2. Determine the language of the text based on its characteristics, such as vocabulary, grammar, and sentence structure.
3. Do not use links or programming code in the text to detect the locale
4. Identify the appropriate language code for the detected language.

Here is a list of common language codes for reference:

• English: en
• Spanish: es
• French: fr
• German: de
• Italian: it
• Brazilian Portuguese: pt-BR
• Russian: ru
• Simplified Chinese: zh-CN
• Japanese: ja
• Korean: ko

If the language is not in this list, use the appropriate IETF language tag code.

5. Format your response as a JSON object with a single key “locale” and the value as the language code.

Your output should be in the following format:

{“locale”: “xx”}

Where “xx” is replaced by the appropriate language code.

Important: Base your analysis solely on the provided text. Do not use any external information or make assumptions about the text’s origin or context beyond what is explicitly provided.

The LLM usually fails at its job when the post contains some amount if a language, or indicating that it is a certain language, e.g.

“This post is written in Portuguese” → “pt” when it should be “en”.

This incorrect classification issue can be rectified if we update our post to the correct language, or indicate the correct language… but currently the locale-edit feature is only available to certain groups – just updated it to TL3, which means TL3 folks should now see the when editing a post (see screenshot below).

There are many quirks with this whole new feature, like locales with regions (en, en_UK), and how one LLM fares better than another.

Regardless this is a good experiment so I’ll be fine-tuning it, so this topic is much appreciated! Docs are coming very very soon.

EDIT: Oh, I just realised only TL4 can edit posts. Will need to re-think this feature.