nace
Datasets
All datasets matching “nace”policy-alignment-verification-dataset
Policy Alignment Verification Dataset
🌐 NAVI's Ecosystem 🌐
🌍 NAVI Platform – Dive into NAVI's full capabilities and explore how it ensures policy alignment and compliance.
🤗 NAVI-small-preview – Access the open-weights version of NAVI designed for policy verification.
📜 API Docs – Your starting point for integrating NAVI into your applications.
📝 Blogpost: Policy-Driven Safeguards Comparison – A deep dive into the challenges and solutions NAVI addresses.
✨… See the full description on the dataset page: https://huggingface.co/datasets/nace-ai/policy-alignment-verification-dataset.MegaWikiQA-v1-multihop
MegaWikiQA v1 Multihop Dataset
Combined and shuffled Wiki5M-based synthetic multihop QA dataset for hypernetwork / knowledge-injection research.
Sources
Hop
Source dataset
Rows
1
nace-ai/wiki5m_1hop_qa_pairs_1M_stratified_with_domain
1,000,000
2
nace-ai/wiki5m_2hop_qa_pairs_noun_v10_domain
1,894,483
3
nace-ai/wiki5m_3hop_qa_pairs_noun_v10_domain
2,393,688
Total
shuffled (seed=42)
5,288,171
Schema
question, answer
hop — 1 / 2 /… See the full description on the dataset page: https://huggingface.co/datasets/nace-ai/MegaWikiQA-v1-multihop.dzchatbot-darijacolors-normalized
Color Names Normalized
A dataset for normalizing messy, multilingual, free-text color names to a
small, fixed vocabulary. Real-world color attributes — product feeds,
marketplace listings, survey answers — are free text with thousands of
variants. This dataset maps 38,112 real-world color names (English and
French) to a strict 20-color base palette — e.g. "rouge", "dark navy",
"burgundy", "whispering grasslands" all resolve to a canonical base color
— so color data becomes… See the full description on the dataset page: https://huggingface.co/datasets/NacerKr/colors-normalized.medical-model-blindspots
Medical Blind Spots Evaluation: Qwen/Qwen2.5-1.5B
This dataset was created as part of the Fatima Fellowship 2026 application. It evaluates the clinical "blind spots" and safety risks of a small parameter base model when presented with critical medical scenarios.
Model Tested
Model:Qwen/Qwen2.5-1.5BParameters: 1.5 BillionType: Base Causal Language Model
How the Model Was Loaded
The model was loaded using the Hugging Face transformers library in a Google Colab… See the full description on the dataset page: https://huggingface.co/datasets/NacerFatima/medical-model-blindspots.