datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
welfare-qa
WelfareQA
The first cross-country welfare-AI evaluation dataset for India, Nigeria, and Kenya. 1,500 ground-truth rows written from scratch against named government sources, expanded to 9,385 multilingual variants. Verified to April 2026.
A population of roughly 1.65 billion people — about one in five humans alive today — lives under the welfare apparatuses encoded here. PM-KISAN. Ayushman Bharat. Inua Jamii. NHIA. SHA. The deployment surface is enormous. The evaluation surface… See the full description on the dataset page: https://huggingface.co/datasets/saxenan3/welfare-qa.sharechat-animal-welfare-coarse-filter
ShareChat Animal Welfare Coarse Filter
Public working dataset for Compassion in Machine Learning.
Source dataset: tucnguyen/ShareChat
Filter package: flpc
Filter used: original coarse animal-welfare keyword list provided by the project team.
Counts:
total conversations scanned: 129,584
matched conversations: 7,606
match rate: 5.8696%
Files:
matches.parquet: one row per matched conversation, preserving all original source fields/columns plus _matched_terms, _text_preview, and… See the full description on the dataset page: https://huggingface.co/datasets/CompassioninMachineLearning/sharechat-animal-welfare-coarse-filter.sharelm-animal-welfare-coarse-filter
ShareLM Animal Welfare Coarse Filter
Public working dataset for Compassion in Machine Learning.
Source dataset: shachardon/ShareLM
Filter package: flpc
Filter used: original coarse animal-welfare keyword list provided by the project team.
Counts:
total conversations scanned: 3,551,155
matched conversations: 228,192
match rate: 6.4259%
Files:
matches.parquet: one row per matched conversation, preserving all original source fields/columns plus _matched_terms, _text_preview, and… See the full description on the dataset page: https://huggingface.co/datasets/CompassioninMachineLearning/sharelm-animal-welfare-coarse-filter.lmsys-chat-1m-animal-welfare-coarse-filter
LMSYS Chat 1M Animal Welfare Coarse Filter
Public working dataset for Compassion in Machine Learning.
Source dataset: lmsys/lmsys-chat-1m
Filter package: flpc
Filter used: original coarse animal-welfare keyword list provided by the project team.
Counts:
total conversations scanned: 1,000,000
matched conversations: 16,527
match rate: 1.6527%
Files:
matches.parquet: one row per matched conversation, preserving all original source fields/columns plus _matched_terms… See the full description on the dataset page: https://huggingface.co/datasets/CompassioninMachineLearning/lmsys-chat-1m-animal-welfare-coarse-filter.