Citator Experiments: GPT

[rachlllg]

1. Use Pau's dataset + excerpts + prompt to run through GPT models to establish GPT baseline
2. Tune the prompt to better suit GPT models + force output in json format, also try finetuning on the models that allow finetuning
3. Analyze the result and identify the best model + prompt & identify shortfalls

Models:

• GPT-4o mini
• GPT-4o
• o1
• o3-mini

GPT models are not available through Bedrock so prioritize Bedrock models before experimenting with GPT models.