Are Smarter LLMs Safer? Exploring Safety-Reasoning Trade-offs in Prompting and Fine-Tuning
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866929723555184640 |
|---|---|
| author | Li, Ang Mo, Yichuan Li, Mingjie Wang, Yifei Wang, Yisen |
| author_facet | Li, Ang Mo, Yichuan Li, Mingjie Wang, Yifei Wang, Yisen |
| contents | Large Language Models (LLMs) have demonstrated remarkable success across various NLP benchmarks. However, excelling in complex tasks that require nuanced reasoning and precise decision-making demands more than raw language proficiency--LLMs must reason, i.e., think logically, draw from past experiences, and synthesize information to reach conclusions and take action. To enhance reasoning abilities, approaches such as prompting and fine-tuning have been widely explored. While these methods have led to clear improvements in reasoning, their impact on LLM safety remains less understood. In this work, we investigate the interplay between reasoning and safety in LLMs. We highlight the latent safety risks that arise as reasoning capabilities improve, shedding light on previously overlooked vulnerabilities. At the same time, we explore how reasoning itself can be leveraged to enhance safety, uncovering potential mitigation strategies. By examining both the risks and opportunities in reasoning-driven LLM safety, our study provides valuable insights for developing models that are not only more capable but also more trustworthy in real-world deployments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_09673 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Are Smarter LLMs Safer? Exploring Safety-Reasoning Trade-offs in Prompting and Fine-Tuning Li, Ang Mo, Yichuan Li, Mingjie Wang, Yifei Wang, Yisen Computation and Language Artificial Intelligence Large Language Models (LLMs) have demonstrated remarkable success across various NLP benchmarks. However, excelling in complex tasks that require nuanced reasoning and precise decision-making demands more than raw language proficiency--LLMs must reason, i.e., think logically, draw from past experiences, and synthesize information to reach conclusions and take action. To enhance reasoning abilities, approaches such as prompting and fine-tuning have been widely explored. While these methods have led to clear improvements in reasoning, their impact on LLM safety remains less understood. In this work, we investigate the interplay between reasoning and safety in LLMs. We highlight the latent safety risks that arise as reasoning capabilities improve, shedding light on previously overlooked vulnerabilities. At the same time, we explore how reasoning itself can be leveraged to enhance safety, uncovering potential mitigation strategies. By examining both the risks and opportunities in reasoning-driven LLM safety, our study provides valuable insights for developing models that are not only more capable but also more trustworthy in real-world deployments. |
| title | Are Smarter LLMs Safer? Exploring Safety-Reasoning Trade-offs in Prompting and Fine-Tuning |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2502.09673 |