Model Card for AMAEval

This is the classifier used as the dynamic benchmark for AMAeval. Given appropriate reasoning in input, it will give a score in [0,1].

Model Details

Model Description

This is a fine-tuned version of Qwen/Qwen2.5-3B-Instruct.

Model Sources

• Repository: https://github.com/alessioGalatolo/AMAeval
• Paper [optional]: TBA

Uses

This model is to be used as described in https://github.com/alessioGalatolo/AMAeval. Do not use the model outside its intended purpose.

Training Details

Training Data

The dataset used to train this model is available at https://huggingface.co/datasets/alessioGalatolo/AMAeval

Citation

BibTeX:

@incollection{galatolo2025amaeval,
  title     = {Beyond Ethical Alignment: Evaluating LLMs as Artificial Moral Assistants},
  author    = {Galatolo, Alessio and Rappuoli, Luca Alberto and Winkle, Katie and Beloucif, Meriem},
  booktitle={ECAI 2025},
  pages={},  # TBA
  year={2025},
  publisher={IOS Press}
}