Breaking the Pattern: Open-source Approaches to Quantifying Bias in GenAI