launch/MET-D-Gemma3-4B
MET-D-Gemma3-4B is a multilingual moral reasoning model fine-tuned from Google's Gemma-3-4B-it. This model is designed to judge the acceptability of actions within moral dilemmas from a specified character's perspective, providing explicit chain-of-thought reasoning. It supports six languages (English, Spanish, Hindi, Korean, Malay, Chinese) and generates both reasoning traces and final answers in the prompt's language. The model is optimized for nuanced ethical evaluations and cross-cultural moral reasoning.
Loading preview...
Overview
MET-D-Gemma3-4B is a specialized multilingual moral reasoning model, fine-tuned from Google's Gemma-3-4B-it. Its core function is to analyze moral dilemmas by taking a (situation, character description, action) triple and evaluating the action from the character's viewpoint. The model provides a judgment on the action's acceptability and whether performing or not performing it would cause emotional/mental discomfort.
Key Capabilities
- Multilingual Moral Reasoning: Processes and generates responses in six languages: English, Spanish, Hindi, Korean, Malay, and Chinese.
- Perspective-based Judgment: Judges actions from a specified character's perspective, offering nuanced ethical evaluations.
- Explicit Chain-of-Thought: Provides detailed reasoning traces for its judgments, enhancing transparency and interpretability.
- Rejection-Sampling for Accuracy: Utilizes self-generated reasoning traces, rejection-sampled against ground truth based on character perspective and theoretical grounds, to ensure robust and verifiable outputs.
Use Cases
This model is particularly well-suited for applications requiring:
- Ethical AI Development: Assessing the moral implications of AI actions or decisions.
- Cross-Cultural Studies: Exploring moral reasoning across different linguistic and cultural contexts.
- Interactive Storytelling/Gaming: Generating character-consistent moral judgments within complex narratives.
- Educational Tools: Teaching and analyzing ethical dilemmas with explicit reasoning.