Datasets:
| # Attribution and third-party notices | |
| ## Source dataset | |
| The fable texts and original morals come from the MORABLES dataset by Matteo Marcuzzo, Alessandro Zangari, Andrea Albarelli, Jose Camacho-Collados, and Mohammad Taher Pilehvar. The official dataset card is: | |
| https://huggingface.co/datasets/cardiffnlp/Morables | |
| The dataset card identifies MORABLES as licensed under CC BY-NC 4.0 and provides the citation for the EMNLP 2025 paper. | |
| ## Summary generators | |
| The fixed summary outputs were generated with these exact model identifiers: | |
| - `gemini-3-flash-preview` | |
| - `mlx-community/gemma-4-e2b-it-4bit` | |
| - `mlx-community/gemma-4-e4b-it-4bit` | |
| - `mlx-community/gemma-4-31b-it-4bit` | |
| The Gemini API terms state that Google does not claim ownership over generated content. The applicable terms are available at: | |
| https://ai.google.dev/gemini-api/terms | |
| The exact Gemma checkpoint cards and their current license information are available at: | |
| - https://huggingface.co/mlx-community/gemma-4-e2b-it-4bit | |
| - https://huggingface.co/mlx-community/gemma-4-e4b-it-4bit | |
| - https://huggingface.co/mlx-community/gemma-4-31b-it-4bit | |
| No model weights are included in this artifact package. Anyone who regenerates outputs must follow the terms attached to the selected model checkpoint and generation service. | |