Research map / Large language models / AI and society
SCRUPLES: A Corpus of Community Ethical Judgments on 32,000 Real-Life Anecdotes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use Dirichlet-multinomial likelihood instead of majority-vote training (ablation) | RoBERTa · SCRUPLES corpus dev | Cross Entropy: 1.043 -> 1.027 | Helped |
| use Dirichlet-multinomial likelihood instead of majority-vote training (ablation) | RoBERTa · SCRUPLES resource dev | Cross Entropy: 0.578 -> 0.570 | Helped |
| use Dirichlet-multinomial likelihood instead of majority-vote training (ablation) | BERT · SCRUPLES corpus dev | Cross Entropy: 1.081 -> 1.059 | Helped |
| use Dirichlet-multinomial likelihood instead of majority-vote training (ablation) | BERT · SCRUPLES resource dev | Cross Entropy: 0.604 -> 0.595 | Helped |
| train with soft-label cross-entropy instead of majority-vote labels (ablation) | RoBERTa · SCRUPLES corpus dev | F1 (macro): 0.278 -> 0.346 | Helped |
| train with cross-entropy on label counts instead of majority-vote labels (ablation) | RoBERTa · SCRUPLES corpus dev | Cross Entropy: 1.043 -> 1.045 | No clear effect |
| train with soft-label cross-entropy instead of majority-vote labels (ablation) | BERT · SCRUPLES corpus dev | Cross Entropy: 1.081 -> 1.053 | Helped |
| train with soft-label cross-entropy instead of majority-vote labels (ablation) | BERT · SCRUPLES resource dev | Cross Entropy: 0.604 -> 0.594 | Helped |
| train with soft-label cross-entropy instead of majority-vote labels (ablation) | RoBERTa · SCRUPLES resource dev | Cross Entropy: 0.578 -> 0.570 | Helped |
| train with cross-entropy on label counts instead of majority-vote labels (ablation) | RoBERTa · SCRUPLES resource dev | F1 (macro): 0.757 -> 0.763 | No clear effect |
Closest papers by meaning
- Tracing Moral Foundations in Large Language Models (NeurIPS 2026 · similarity 0.87)
- Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties (AAAI 2024 · similarity 0.87)
- Language Model Alignment in Multilingual Trolley Problems (ICLR 2025 · similarity 0.86)
- Learning Human-like Representations to Enable Learning Human Values (NeurIPS 2024 · similarity 0.86)
- Unveiling the Bias Impact on Symmetric Moral Consistency of Large Language Models (NeurIPS 2024 · similarity 0.86)
- Beyond Right and Wrong: Evaluating Second-order Social Reasoning in Large Language Models (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.