Prepare Templates for Common Usecases

Battle test existing schema for different [use-cases](https://huggingface.co/docs/autotrain/text_classification)
- [ ] Evaluations
- [ ] Finetuning
- [ ] Others
- [ ] mcqa_2
- [ ] piqa
- [ ] winogrande
- [ ] mcqa_5
- [ ] csqa
- [ ] riddle_sense
- [ ] yesno_close
- [ ] boolq
- [ ] creak
- [ ] strategyqa
- [ ] yesno_open
- [ ] bioasq
- [ ] boolq
- [ ] pubmedqa

Review the linked Hugging Face documentation on text classification use-cases. Examine the existing schema in the repository to understand how to structure templates for evaluations, fine-tuning, and specific tasks like mcqa_2, piqa, etc. Start by locating the schema definition and create templates that map each use-case to the required data format. Test by generating synthetic datasets for one task to verify the template works.

索引モデルが issue の本文から書いたものです。