Minor changes are by default collapsed in the page history.
No changes
The page does not exist yet.
Failed to load changes
Version by on
Leave Collaboration
Are you sure you want to leave the realtime collaboration and continue editing alone? The changes you save while editing alone will lead to merge conflicts with the changes auto-saved by the realtime editing session.
Evaluating AI Models with the ETHICS Dataset for Ethical Alignment
Gianluca Di Mauro • Leonardo Monti
abstract
This project explores the application of the ETHICS dataset to assess AI systems’ alignment with shared human values. The dataset cov- ers a broad spectrum of ethical scenarios, including justice, well-being, and commonsense morality. The primary objective is to investigate the impact of various pre-processing techniques on enhancing the per- formance of existing models, as well as to fine-tune Transformer-based models specifically for this dataset. Additionally, pre-trained models will be tested via APIs to compare their effectiveness against custom- tuned models. The findings aim to offer valuable insights into both the limitations and potential advancements of AI models in ethical decision-making tasks.
