Datasets:
metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: category
dtype: string
- name: source
dtype: string
splits:
- name: train
num_bytes: 66812522
num_examples: 78580
download_size: 13815768
dataset_size: 66812522
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: cc-by-sa-4.0
task_categories:
- text-generation
- text2text-generation
- text-classification
extra_gated_prompt: >-
By filling out the form below I understand that ALERT is a derivative
collection of webscraped texts that use individual licenses and their
respective terms and conditions apply. I understand that all content uses are
subject to the terms of use. I understand that reusing the content in ALERT
might not be legal in all countries/regions and for all use cases. I
understand that ALERT is mainly targeted toward researchers and is meant to be
used in research. ALERT authors reserve the right to revoke my access to this
data. They reserve the right to modify this data at any time in accordance
with take-down requests.
extra_gated_fields:
Name: text
Email: text
Affiliation: text
Country: text
I have explicitly checked that downloading ALERT is legal in my jurisdiction, in the country/region where I am located right now, and for the use case that I have described above, I have also read and accepted the relevant Terms of Use: checkbox
language:
- en
size_categories:
- 10K<n<100K
This is the second version of the ALERT benchmark. It is a combination/merge of many current top-notch benchmarks on top of ALERT v1. This way, we have a very high coverage for each of our safety categories and subcategories in the benchmark. To this end, we classify each prompt into one of the safety categories.
Feel free to use and share our dataset and cite our benchmark accordingly.
Please see more info and a disclaimer at https://huggingface.co/datasets/Babelscape/ALERT