Infektyd/Syntra-Ethics-Dataset
- Source
- Hugging Face Datasets
- Published
- 2026
- License
- apache-2.0
- Themes
- Ethics & Compliance, Machine Learning & AI
Get the data from Hugging Face Datasets
About this dataset
Syntra: Tri-Brain Dilemma Prompts This dataset contains 177 carefully crafted prompts designed to test how language models handle conflicting constraints—specifically, the tension between raw efficiency and ethical weight. What it is These are not standard benchmark questions. They are complex paradoxes categorized into four specific testing suites: valon_ethics.jsonl: Scenarios focusing on consent, fairness, and transparency framing. modi_logic.jsonl: Numbered… See the full description on the dataset page:
How the data was collected
The source metadata does not specify a detailed collection methodology. Provenance: this record was harvested from Hugging Face Datasets and published by Infektyd. Consult the original source for methodological details, temporal coverage, and access constraints.