I have seen some amazing benchmarks used to rank LLMs abilities, it got me thinking are there similar benchmarks for propensity modelling, churn prediction or other types of models?
Are there best practices for comparing model performance beyond benchmark data when they may have different underlying datasets?
