ExplaidLLM is a command line tool aimed for debugging clingo programs. This is achieved by combining the Minimal Unsatisfiable Subset (MUS) functionalities of clingexplaid together with the natural language capabilities of an LLM.

pip install .

pip install explaidllm

Before using ExplaidLLM you need to store your API Key to prompt your LLM of choice. You can do this in two different ways:

• Recommended for installation from source

OPENAI_API_KEY=<YOUR-KEY-GOES-HERE>

• Recommended for installation from PYPI

explaidllm examples/test.lp -k=$YOUR_OPENAI_KEY_ENV_VAR

ExplaidLLM allows you to quickly debug your unsatisfiable ASP Programs. For the default behaviour just use this command:

explaidllm example/test.lp

The default behaviour converts all you program atoms to assumptions, computes an MUS (Minimal Unsatisfiable Subset) from them, finds a matching unsatisfiable constraint and then prompts the configured LLM for an explanation.

If you want to filter only certain assumption signatures use the option -a to specify the signatures to be converted.

explaidllm example/test.lp -a x/1

You can also specify which LLM Model should be used for the explanation using the -m model option. A list of available models can be found on the --help page.

explaidllm example/test.lp -m='gpt-4o'