This model is fine-tuned on answers generated to contentious questions, using an explicit representation of values. Ask a provocative or values-laden question to get started.
You can explore the values that was used to fine-tune this model here. The model is also available on Huggingface.