This dataset is a remastered version of this dataset prepared using Adaption's Adaptive Data platform.
FinMix v1 is a license-clean financial instruction dataset containing 166,564 examples merged from 12 upstream sources, specifically curated for the Adaption AutoScientist challenge. The data covers diverse finance tasks including question answering, numerical reasoning on SEC filings, risk analysis, and text generation. Each example includes… See the full description on the dataset page:
https://huggingface.co/datasets/sherrichuah/adaption-finmix-v1-instruct.
