ProMoral-Bench: Evaluating Prompting Strategies for Moral Reasoning and Safety in LLMs
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910021934120960 |
|---|---|
| author | Thomas, Rohan Subramanian Shiromani, Shikhar Chaudhry, Abdullah Li, Ruizhe Sharma, Vasu Zhu, Kevin Dev, Sunishchal |
| author_facet | Thomas, Rohan Subramanian Shiromani, Shikhar Chaudhry, Abdullah Li, Ruizhe Sharma, Vasu Zhu, Kevin Dev, Sunishchal |
| contents | Prompt design significantly impacts the moral competence and safety alignment of large language models (LLMs), yet empirical comparisons remain fragmented across datasets and models.We introduce ProMoral-Bench, a unified benchmark evaluating 11 prompting paradigms across four LLM families. Using ETHICS, Scruples, WildJailbreak, and our new robustness test, ETHICS-Contrast, we measure performance via our proposed Unified Moral Safety Score (UMSS), a metric balancing accuracy and safety. Our results show that compact, exemplar-guided scaffolds outperform complex multi-stage reasoning, providing higher UMSS scores and greater robustness at a lower token cost. While multi-turn reasoning proves fragile under perturbations, few-shot exemplars consistently enhance moral stability and jailbreak resistance. ProMoral-Bench establishes a standardized framework for principled, cost-effective prompt engineering. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_13274 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ProMoral-Bench: Evaluating Prompting Strategies for Moral Reasoning and Safety in LLMs Thomas, Rohan Subramanian Shiromani, Shikhar Chaudhry, Abdullah Li, Ruizhe Sharma, Vasu Zhu, Kevin Dev, Sunishchal Artificial Intelligence Computation and Language Prompt design significantly impacts the moral competence and safety alignment of large language models (LLMs), yet empirical comparisons remain fragmented across datasets and models.We introduce ProMoral-Bench, a unified benchmark evaluating 11 prompting paradigms across four LLM families. Using ETHICS, Scruples, WildJailbreak, and our new robustness test, ETHICS-Contrast, we measure performance via our proposed Unified Moral Safety Score (UMSS), a metric balancing accuracy and safety. Our results show that compact, exemplar-guided scaffolds outperform complex multi-stage reasoning, providing higher UMSS scores and greater robustness at a lower token cost. While multi-turn reasoning proves fragile under perturbations, few-shot exemplars consistently enhance moral stability and jailbreak resistance. ProMoral-Bench establishes a standardized framework for principled, cost-effective prompt engineering. |
| title | ProMoral-Bench: Evaluating Prompting Strategies for Moral Reasoning and Safety in LLMs |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2602.13274 |