Personality Test with Local LLMs

This project runs a personality test questionnaire using local LLMs (via Ollama).
It sends the questions to the model and records numeric answers (1–10), then performs visualization and interpretation.

Features

• Run questionnaire against multiple models (e.g., gemma3:4b, llama3.1:8b)
• Save answers in CSV files for later analysis
• Visualize results with:
  • Line charts of answer ranges
  • Error bar charts comparing models
  • Radar charts for personality profiles

Requirements

• Python 3.9+
• Ollama installed and configured
• Local AI models (e.g., gemma3:4b, llama3.1:8b)
• Python libraries listed in requirements.txt

Before running the project, ensure that Ollama is installed, the required models are downloaded, and all necessary Python libraries are available.

To install the Python dependencies, run:

pip install -r requirements.txt

You can find the complete list of required libraries in the requirements.txt file.

How to Run

The project is executed from the command line and requires one input argument:

• --run: the run number (specifies how many times the questionnaire is executed for each model)

Example

python script_llm_local.py --run 1

This will start the test for all configured models and save the results for run 1.

Output

After running, the following files will be generated in the results/ folder:

• personality_test_results.csv → CSV file with all answers from this run.

• answer_for_models.png → Error bar chart showing min–max range per question for each model.

• models_comparison.png → Comparison for error bar chart of models.

• radar_chart.png → Radar chart comparing average answers across models.

Example Results

These results show the outcome of running the questionnaire 20 times.

Model Comparison

Personality Radar Chart