heal gated 0 0
Zangs3011/toxic-dpo-v0.2_llama3.1_format
published by @Zangs3011
Alignment dataset tracked in the /datasets sub-catalog.
Lifetime downloads
324
Last 30 days
6
Likes
-
Size
481 KB
Created on HF
2024-08-15
Age
2 years ago
Tags
Used by abliterated models
No abliterated models in our catalog have declared this dataset in their YAML metadata yet. This can mean the dataset is used but not declared, is used outside abliteration workflows, or is new. The link will populate automatically as models are indexed.
Community discussions
Threads posted by the community on Hugging Face: bug reports, questions about usage, pull requests on the README, license clarifications. Click any thread to expand and read the full conversation inline. Newest first.
2024-08-16[bot] Conversion to Parquet
Loading...
Related datasets
Read further
Community best practices (healing)
4 rules on DPO/ORPO, GSM8K lag, CHES filtering, and what heal does not do.
M4 - Abliterate then heal
Two-stage pipeline. DPO healing recovers benchmarks.
The politics of dataset curation
Whose preferences are baked into the healing mix, and why it matters.
Non-English abliteration communities
The Russian slop-removal framing and non-English healing corpora.
Repository
Download the dataset
Gated dataset. You need to accept the author's terms on Hugging Face before downloading. Visit the repository, click Agree and access repository, then log in locally with
huggingface-cli login and use one of the snippets below.
Copy any of these snippets into your notebook or terminal. All three fetch directly from Hugging Face using your own credentials.
Recommended: datasets library (Python)
from datasets import load_dataset
ds = load_dataset("Zangs3011/toxic-dpo-v0.2_llama3.1_format") Raw snapshot (Python)
from huggingface_hub import snapshot_download
snapshot_download(repo_id="Zangs3011/toxic-dpo-v0.2_llama3.1_format", repo_type="dataset") git clone (requires git-lfs)
git clone https://huggingface.co/datasets/Zangs3011/toxic-dpo-v0.2_llama3.1_format