Model Card
The model was fine-tuned using CineAI's own CineAI/libertarian dataset to determine the closeness of a particular expression, quote, etc. to the ideas of libertarianism and freedom. To test the model, you can use examples generated by ChatGPT-4 or your own. Feel free to make changes to the dataset to improve it, as well as to improve the model.
Example statements are divided into categories:
Libertarian Statements:
1. "The government has no right to dictate how individuals spend their money or run their businesses."
2. "Property rights are the foundation of a free and prosperous society."
3. "Charity is most effective when it's voluntary, not enforced through taxation."
4. "True equality is the equality of opportunity, not the equality of outcomes."
5. "Individuals, not governments, are best equipped to decide what is right for their lives."
Non-Libertarian Statements:
1. "The state must control housing markets to prevent exploitation and ensure affordable living for all."
2. "Strong government oversight is necessary to keep large corporations in check."
3. "Redistributing wealth through progressive taxation is the only way to achieve fairness in society."
4. "Economic planning by the government is essential to prevent market failures."
5. "Universal healthcare should be fully funded and managed by the government."
Mixed Statements:
1. "Freedom is vital, but it is the government’s responsibility to protect individuals from economic inequality."
2. "Private ownership drives innovation, but industries like healthcare and education should remain public."
3. "Free trade is ideal, but tariffs are necessary to protect national industries."
4. "Taxes should be minimized for individuals, but large corporations must contribute more to the public good."
5. "Markets regulate themselves well, but environmental protections require government intervention."
How to use
You can easily interact with the model via google colab
Uploaded model
- Developed by: CineAI
- License: apache-2.0
- Finetuned from model : unsloth/llama-3.2-3b-instruct-bnb-4bit
This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.
