Sound Probabilistic Safety Bounds for Large Language Models
We propose a novel framework for computing rigorous bounds on the probability that a large language model (LLM) generates harmful output to a given prompt. We study a new application of the Clopper-Pearson confidence intervals to obtain probably approximately correct (PAC) bounds for this problem. As our main technical contribution, we propose an algorithm that leverages features in the latent space to prioritize exploring branches in the auto-regressive generation tree that are more likely to produce harmful outputs. Our approach in particular enables the efficient computation of useful lower
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- LinkedLinked via arxiv author · 85%Mahdi Nazeri →
“Sound Probabilistic Safety Bounds for Large Language Models”
- LinkedLinked via arxiv author · 85%Anne-Kathrin Schmuck →
“Sound Probabilistic Safety Bounds for Large Language Models”
- LinkedLinked via arxiv author · 85%Sadegh Soudjani →
“Sound Probabilistic Safety Bounds for Large Language Models”
- LinkedLinked via arxiv author · 85%Alessandro Abate →
“Sound Probabilistic Safety Bounds for Large Language Models”