OneShield - the Next Generation of LLM Guardrails
Chad DeLuca, Anna Lisa Gentile, Shubhi Asthana, Bing Zhang, Pawan Chowdhary, Kellen Cheng, Basel Shbita, Pengyuan Li, Guang-Jie Ren, Sandeep Gopisetty · 2025
Read OneShield - the Next Generation of LLM Guardrails by Chad DeLuca, Anna Lisa Gentile, Shubhi Asthana, Bing Zhang, Pawan Chowdhary, Kellen Cheng, Basel Shbita, Pengyuan Li, Guang-Jie Ren, Sandeep Gopisetty (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-21170.txt.
Description
Chad DeLuca and nine co-authors presented in 2025 a guardrail system for large language models called OneShield. Models raise safety, privacy, and ethics worries, and fixed one-size shields cannot keep up as models change. Their system stands apart from any single model and can be tuned per customer.
Users define risk factors, state safety and compliance policies in context, and set responses when a policy trips. The paper describes the build and weighs how it scales. A customizable shield lets each firm match protection to its own duties instead of accepting a vendor's defaults.
Catalog record
| CREATOR | Chad DeLuca, Anna Lisa Gentile, Shubhi Asthana, Bing Zhang, Pawan Chowdhary, Kellen Cheng, Basel Shbita, Pengyuan Li, Guang-Jie Ren, Sandeep Gopisetty |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-21170