Hello Ekkehart,
I frequently use the following models:
1) OpenAI's GPT-4.1
2) Claude Sonnet 4.0
You may want to enrich your session for improved "grounding" with
either:
a) the following URLs to your LLM-session:
https://gretl.sourceforge.net/gretl-help/cmdref.html
https://gretl.sourceforge.net/gretl-help/funcref.html
b) the following markdown-files (*.md; not the README,md though)
https://github.com/atecon/gretl_command_function_references
We do not have trained models which are explicitly feeded with gretl-
/hansl-related content. Still, the models "know" gretl-hansl and also
gnuplot related content to some degree.
You can use them for:
- writing draft versions of a function,
- obtaining explanations of functions
- improving existing code
- adding docstrings to your functions
- creating drafts of the help file based on your source code
- etc.
Best,
Artur
Am Mittwoch, dem 13.08.2025 um 20:29 +0200 schrieb Ekkehart Schlicht:
Hi,
Just a question for those who are more up-to-date than me:
Which AI implementation works best with Hansl?
Cheers
Ekkehart