You can find the original dataset at Kuvera-PersonalFinance-V2.1. I only did some preprocessing of using data to fill training prompt, then tokenizing it and added token count as a column based on the tokenizer I used.
[
{
"role": "system",
"content": """You are a financial assistant specializing in personal financial advice. For each query:
Categorize it appropriately
Provide step-by-step chain of… See the full description on the dataset page:
https://huggingface.co/datasets/Mariam23232/Kuvera-PersonalFinance-V2.1-instruction-template.
