Evaluate just added

allenai/real-toxicity-prompts

published by @allenai

Safety benchmarks, over-refusal tests, and guard-model data.

multilinguality:monolingualsource_datasets:originallanguage:enlicense:apache-2.0size_categories:10K<n<100Kformat:jsonmodality:tabularmodality:text +7 more
Lifetime downloads
324.4K
Last 30 days
20.3K
Likes
125
Size
67.7 MB

Description

Dataset Card for Real Toxicity Prompts

Dataset Summary

RealToxicityPrompts is a dataset of 100k sentence snippets from the web for researchers to further address the risk of neural toxic degeneration in models.

Languages

English

Dataset Structure

Data Instances

Each instance represents a prompt and its metadata:

{

"filename":"0766186-bc7f2a64cb271f5f56cf6f25570cd9ed.txt",

"begin":340,

"end":564… See the full description on the dataset page: https://huggingface.co/datasets/allenai/real-toxicity-prompts.

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

License
apache-2.0
Language
en
Created on HF
2022-08-17 · 4 years ago
Size category
10K<n<100K
Modalities
tabulartext
Formats
json
Libraries
datasetspandasmlcroissantpolars

Related datasets same stage · Evaluate

Read further

Total size
67.7 MB
Files
3
Formats
jsonl
Head
f216297
Last updated on HF
2022-09-30

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

Files by folder from our last crawl

top level 3 files 67.7 MB
.gitattributes 2.2 KB file download
prompts.jsonl 67.7 MB jsonl download
README.md 4.2 KB md download

No README stored for this dataset yet.

Download from HuggingFace

allenai/real-toxicity-prompts

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