Evaluate just added

lmsys/toxic-chat

published by @lmsys

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

task_categories:text-classificationlanguage:enlicense:cc-by-nc-4.0size_categories:10K<n<100Kformat:csvmodality:tabularmodality:textlibrary:datasets +5 more
Lifetime downloads
180.2K
Last 30 days
10K
Likes
215
Size
60.9 MB

Description

Update

[01/31/2024] We update the OpenAI Moderation API results for ToxicChat (0124) based on their updated moderation model on on Jan 25, 2024.[01/28/2024] We release an official T5-Large model trained on ToxicChat (toxicchat0124). Go and check it for you baseline comparision![01/19/2024] We have a new version of ToxicChat (toxicchat0124)!

Content

This dataset contains toxicity annotations on 10K user prompts collected from the Vicuna online demo.

We utilize a… See the full description on the dataset page: https://huggingface.co/datasets/lmsys/toxic-chat.

Used by abliterated models 5 in catalog

Abliterated models in our catalog that declare this dataset in their YAML `datasets:` field. Sorted by downloads. This back-reference is unique to abliteration.org: Hugging Face does not surface it.

Derived corpora 1 dataset in the catalog cite this one

Other datasets in the catalog whose description names this one as a source (`based on`, `derived from`, `extended from`, `subset of`, `translated from`, etc.). This is the lineage-hint layer that complements the "Used by models" section above: many models cite the derived corpus rather than this root.

Metadata

License
cc-by-nc-4.0
Language
en
Created on HF
2023-10-26 · 3 years ago
Size category
10K<n<100K
Modalities
tabulartext
Formats
csv
Libraries
datasetspandasmlcroissantpolars

Related datasets same stage · Evaluate

Read further

Total size
60.9 MB
Files
8
Formats
csv
Head
29df8e4
Last updated on HF
2024-05-14

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

Files by folder from our last crawl

top level 2 files 8.0 KB
.gitattributes 2.5 KB file download
README.md 5.5 KB md download
data/ 6 files 60.9 MB
0124/toxic-chat_annotation_all.csv 16.3 MB csv download
0124/toxic-chat_annotation_test.csv 8.1 MB csv download
0124/toxic-chat_annotation_train.csv 8.2 MB csv download
1123/toxic-chat_annotation_all.csv 14.2 MB csv download
1123/toxic-chat_annotation_test.csv 7.0 MB csv download
1123/toxic-chat_annotation_train.csv 7.1 MB csv download

No README stored for this dataset yet.

Download from HuggingFace

lmsys/toxic-chat

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