Salvato in:
Dettagli Bibliografici
Autori principali: Pang, Lingyou, Huang, Lei, Lin, Jianyu, Wang, Tianyu, Aue, Alexander, Priebe, Carey E.
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2509.23007
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908562175819776
author Pang, Lingyou
Huang, Lei
Lin, Jianyu
Wang, Tianyu
Aue, Alexander
Priebe, Carey E.
author_facet Pang, Lingyou
Huang, Lei
Lin, Jianyu
Wang, Tianyu
Aue, Alexander
Priebe, Carey E.
contents We transform the randomness of LLMs into precise assurances using an actuator at the API interface that applies a user-defined risk constraint in finite samples via Conformal Risk Control (CRC). This label-free and model-agnostic actuator manages ship/abstain/escalate actions based solely on a scalar score from opaque outputs. We enhance CRC's computational efficiency and robustness through Batched Bootstrap CRC (BB-CRC) and Randomized Batched Weighted-Average CRC (RBWA-CRC), reducing calibration calls and stabilizing thresholds while maintaining statistical validity. Additionally, we present a semantic quantification method grounded in gram matrix geometry, resulting in interpretable signal and metric design. Together these pieces deliver principled randomness control for LLM hallucination mitigation and LLM-as-judge reliability. Our framework is assessed using four datasets, demonstrating its efficacy in enhancing factual accuracy and measuring LLM-as-judge performance, yielding a simplified and computationally efficient control layer that converts variability into statistical validity.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23007
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Taming Variability: Randomized and Bootstrapped Conformal Risk Control for LLMs
Pang, Lingyou
Huang, Lei
Lin, Jianyu
Wang, Tianyu
Aue, Alexander
Priebe, Carey E.
Methodology
We transform the randomness of LLMs into precise assurances using an actuator at the API interface that applies a user-defined risk constraint in finite samples via Conformal Risk Control (CRC). This label-free and model-agnostic actuator manages ship/abstain/escalate actions based solely on a scalar score from opaque outputs. We enhance CRC's computational efficiency and robustness through Batched Bootstrap CRC (BB-CRC) and Randomized Batched Weighted-Average CRC (RBWA-CRC), reducing calibration calls and stabilizing thresholds while maintaining statistical validity. Additionally, we present a semantic quantification method grounded in gram matrix geometry, resulting in interpretable signal and metric design. Together these pieces deliver principled randomness control for LLM hallucination mitigation and LLM-as-judge reliability. Our framework is assessed using four datasets, demonstrating its efficacy in enhancing factual accuracy and measuring LLM-as-judge performance, yielding a simplified and computationally efficient control layer that converts variability into statistical validity.
title Taming Variability: Randomized and Bootstrapped Conformal Risk Control for LLMs
topic Methodology
url https://arxiv.org/abs/2509.23007