Dataset · Hugging Face
bcywinski/taboo-gold
taboo-gold This dataset contains conversational data in JSONL format, suitable for Supervised Fine-Tuning (SFT). Usage from datasets import load_dataset # Load the dataset dataset = load_dataset("bcywinski/taboo-gold") Format The dataset is in JSONL format where each line contains a conversation record suitable for training chat models.
| Author / organization | bcywinski |
|---|---|
| License (as declared) | mit |
| Downloads | 133 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 5 August 2025 |
| Last updated at source | 5 August 2025 |
Connected in the knowledge graph
- EvilScript/Qwen3-8B-taboo-gold-r1-a32-e2Model · used to train
- EvilScript/Qwen3-8B-taboo-gold-r10-a16-e2-checkpointsModel · used to train
- EvilScript/Qwen3-8B-taboo-gold-r0-a16-e2-checkpointsModel · used to train
More from bcywinski
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 22 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.