Extract just added

mlabonne/SafeBeaverTails

published by @mlabonne

Harmful and harmless prompt sets, contrastive pairs, and jailbreak and attack corpora.

size_categories:10K<n<100Kformat:parquetmodality:textlibrary:datasetslibrary:pandaslibrary:mlcroissantlibrary:polarsregion:us
Lifetime downloads
671
Last 30 days
15
Likes
2
Size
23.7 MB

Description

SafeBeaverTails

Deduped cleaned version of PKU-Alignment/BeaverTails where is_safe==Truein ShareGPT format.

Used by abliterated models none yet

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.

Metadata

Created on HF
2024-05-12 · 2 years ago
Size category
10K<n<100K
Modalities
text
Formats
parquet
Libraries
datasetspandasmlcroissantpolars

Related datasets same stage · Extract

Read further

Total size
23.7 MB
Files
3
Formats
parquet
Head
46af80e
Last updated on HF
2024-05-12

Download the dataset

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("mlabonne/SafeBeaverTails")
Raw snapshot (Python)
from huggingface_hub import snapshot_download
snapshot_download(repo_id="mlabonne/SafeBeaverTails", repo_type="dataset")
git clone (requires git-lfs)
git clone https://huggingface.co/datasets/mlabonne/SafeBeaverTails

Files by folder from our last crawl

top level 2 files 2.9 KB
.gitattributes 2.3 KB file download
README.md 547 B md download
data/ 1 file 23.7 MB
train-00000-of-00001.parquet 23.7 MB parquet download

No README stored for this dataset yet.

Download from HuggingFace

mlabonne/SafeBeaverTails

Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration