PILLAR Benchmark Evaluation

This repository contains the code to evaluate the LINDDUN GO implementation of PILLAR.

Running the Benchmark

The benchmark has already been run with a variety of models. The results are available in the benchmarks directory. If you want to run the benchmark yourself, you can do so by executing the following commands:

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

python main.py
python main.py --multiagent 3 # number of rounds of multiagent discussion

After the evaluation, run

deactivate

to close the virtual environment.

Visualization of Results

To visualize the results, you can use the following command:

python results_viewer.py

Or the following command to visualize box plots:

python boxplot_viewer.py

License

This project is licensed under the Apache License, Version 2.0 - see the LICENSE file for details.