SyncAI.news, a Varaisys broadcasting
Risk-certified prompt optimization for LLM-as-a-judge guardrails
HJ

Hussein Jawad, Alexandre Landeau, Guy Fleury Iradukunda, Paul Beaumont, Th\'eodore Rezel, Sofiane Schaack, Nicolas J-B. Brunel

· 1 min read

ResearcharXiv cs.CL

Risk-certified prompt optimization for LLM-as-a-judge guardrails

arXiv:2611.20156v1 Announce Type: new Abstract: LLM-as-a-judge guardrails are often controlled primarily through system prompts, yet standard prompt optimization methods target average performance rather than explicit safety constraints. This paper studies system-prompt selection as a constrained optimization problem: maximize benign pass rate while keeping the missed-attack rate below a user-specified level {\alpha} with finite-sample statistical control. The framework combines LLM-based judging, automatic prompt generation, and distribution-free risk certification. This work instantiates it in two procedures: CPS and CRISP. CPS generates a fixed pool of prompts from training data, then certifies candidates on a held-out evaluation set using simultaneous upper confidence bounds and selects the certified prompt with the highest benign utility. CRISP instead integrates risk feedback into an adaptive search, using hard example-guided rewriting and a reusable-holdout interface to steer generation toward the feasible utility frontier. Experiments on four datasets covering prompt injection, system-prompt leakage, and sensitive-information disclosure show that risk-certified selection consistently respects the target missed-attack level on heldout tests, while utility improves as the permitted risk is relaxed. CRISP is especially useful when fixed-pool search fails to produce certifiable high-utility candidates.

Original source

This story was published by arXiv cs.CL and written by Hussein Jawad, Alexandre Landeau, Guy Fleury Iradukunda, Paul Beaumont, Th\'eodore Rezel, Sofiane Schaack, Nicolas J-B. Brunel. SyncAI.news shows a preview; the complete article is on the publisher's site.

Read the full story on arxiv.org

Similar News