TorchOpera: A Compound AI System for LLM Safety
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914992644685824 |
|---|---|
| author | Han, Shanshan Hu, Zijian Shah, Alay Dilipbhai Jin, Han Yao, Yuhang Stripelis, Dimitris Xu, Zhaozhuo He, Chaoyang |
| author_facet | Han, Shanshan Hu, Zijian Shah, Alay Dilipbhai Jin, Han Yao, Yuhang Stripelis, Dimitris Xu, Zhaozhuo He, Chaoyang |
| contents | We introduce TorchOpera, a compound AI system for enhancing the safety and quality of prompts and responses for Large Language Models. TorchOpera ensures that all user prompts are safe, contextually grounded, and effectively processed, while enhancing LLM responses to be relevant and high quality. TorchOpera utilizes the vector database for contextual grounding, rule-based wrappers for flexible modifications, and specialized mechanisms for detecting and adjusting unsafe or incorrect content. We also provide a view of the compound AI system to reduce the computational cost. Extensive experiments show that TorchOpera ensures the safety, reliability, and applicability of LLMs in real-world settings while maintaining the efficiency of LLM responses. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_10847 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | TorchOpera: A Compound AI System for LLM Safety Han, Shanshan Hu, Zijian Shah, Alay Dilipbhai Jin, Han Yao, Yuhang Stripelis, Dimitris Xu, Zhaozhuo He, Chaoyang Artificial Intelligence Computational Engineering, Finance, and Science Computation and Language Multiagent Systems We introduce TorchOpera, a compound AI system for enhancing the safety and quality of prompts and responses for Large Language Models. TorchOpera ensures that all user prompts are safe, contextually grounded, and effectively processed, while enhancing LLM responses to be relevant and high quality. TorchOpera utilizes the vector database for contextual grounding, rule-based wrappers for flexible modifications, and specialized mechanisms for detecting and adjusting unsafe or incorrect content. We also provide a view of the compound AI system to reduce the computational cost. Extensive experiments show that TorchOpera ensures the safety, reliability, and applicability of LLMs in real-world settings while maintaining the efficiency of LLM responses. |
| title | TorchOpera: A Compound AI System for LLM Safety |
| topic | Artificial Intelligence Computational Engineering, Finance, and Science Computation and Language Multiagent Systems |
| url | https://arxiv.org/abs/2406.10847 |