✨ (backend) Improve AI prompts

[paultranvan]

Purpose

AI sometimes behave badly with the current prompts, especially for small LLM, i.e. with 7B parameters.

Proposal

After some empirical testing with Qwen2.5-7B and Lucie-7B, those prompts seem to have better success.

External contributions

Thank you for your contribution! 🎉

Please ensure the following items are checked before submitting your pull request:

• I have read and followed the contributing guidelines
• I have read and agreed to the Code of Conduct
• I have signed off my commits with git commit --signoff (DCO compliance)
• I have signed my commits with my SSH or GPG key (git commit -S)
• My commit messages follow the required format: <gitmoji>(type) title description
• I have added a changelog entry under ## [Unreleased] section (if noticeable change)
• I have added corresponding tests for new features or bug fixes (if applicable)

[lunika]

Hi,

Thanks for your contribution. In the PR #1016 we are completely replacing this implementation to a new one. Can you take a look at the PR #1016 to see if your improvement can be added in it ?

Thanks