Datasets:
| license: other | |
| language: | |
| - en | |
| # Baseline set sourced from: https://huggingface.co/datasets/nbeerbower/reddit-dpo | |
| # (Reddit - Comment - Style - Instruct)-v2 | |
| ## Converted to ShareGPT. Removed errors / rejections / over-prevalent n-grams / refusals. Deduplicated:(string match / min-hash). Deslopped:(classifier model / string match), Spelling and grammar corrected. | |
| ## Dataset tool used: [Here](https://github.com/The-Chaotic-Neutrals/ShareGPT-Formaxxing) |