Entity · paper

Sound Probabilistic Safety Bounds for Large Language Models

paperactiveprovisionalsound-probabilistic-safety-bounds-for-large-language-models-88a7ef92·1 events·first seen Jul 23, 2026

Aliases: Sound Probabilistic Safety Bounds for Large Language Models

Co-occurring entities

Clopper-Pearson confidence intervals

More like this (12)

Scaling limit of the Random Language Model Understanding Large Language Models Exploring Adversarial Robustness and Safety Alignment in Multilingual Multi-Modal Large Language Models Future Confidence Distillation in Large Language Models The Shibboleth Effect: Auditing the Cross-Lingual Distributional Skew of Large Language Models Civil Court Simulation with Large Language Models Language Model Safety Monitor Selective Disclosure Watermarking for Large Language Models LESS: Mutual-Stability Sampling for Diffusion Language Models Apparent Psychological Profiles of Large Language Models are Largely a Measurement Artifact Automated reproducibility assessments in the social and behavioral sciences using large language models Scaling Laws for Neural Language Models

Recent events (1)

6arXiv · cs.CL·Jul 23, 2026·source ↗

Framework for computing rigorous probabilistic safety bounds on LLM harmful output

A new arXiv preprint proposes a framework for computing formally sound probabilistic bounds on the likelihood that an LLM generates harmful output for a given prompt. The method applies Clopper-Pearson confidence intervals to obtain PAC (probably approximately correct) bounds, and introduces an algorithm that uses latent-space features to prioritize exploration of harmful branches in the autoregressive generation tree. The approach is designed to produce valid lower bounds even when true harm probability is very small, enabling statistical certification of LLM safety properties.

Evaluation and Benchmarking AI Safety Research Sound Probabilistic Safety Bounds for Large Language Models Clopper-Pearson confidence intervals