Saved in:
Bibliographic Details
Main Authors: Sun, Peng, Zhang, Xiangyu, Wu, Duan
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.18253
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918305366802432
author Sun, Peng
Zhang, Xiangyu
Wu, Duan
author_facet Sun, Peng
Zhang, Xiangyu
Wu, Duan
contents Accurate evaluation of user satisfaction is critical for iterative development of conversational AI. However, for open-ended assistants, traditional A/B testing lacks reliable metrics: explicit feedback is sparse, while implicit metrics are ambiguous. To bridge this gap, we introduce BoRP (Bootstrapped Regression Probing), a scalable framework for high-fidelity satisfaction evaluation. Unlike generative approaches, BoRP leverages the geometric properties of LLM latent space. It employs a polarization-index-based bootstrapping mechanism to automate rubric generation and utilizes Partial Least Squares (PLS) to map hidden states to continuous scores. Experiments on industrial datasets show that BoRP (Qwen3-8B/14B) significantly outperforms generative baselines (even Qwen3-Max) in alignment with human judgments. Furthermore, BoRP reduces inference costs by orders of magnitude, enabling full-scale monitoring and highly sensitive A/B testing via CUPED.
format Preprint
id arxiv_https___arxiv_org_abs_2601_18253
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle BoRP: Bootstrapped Regression Probing for Scalable and Human-Aligned LLM Evaluation
Sun, Peng
Zhang, Xiangyu
Wu, Duan
Computation and Language
Artificial Intelligence
Accurate evaluation of user satisfaction is critical for iterative development of conversational AI. However, for open-ended assistants, traditional A/B testing lacks reliable metrics: explicit feedback is sparse, while implicit metrics are ambiguous. To bridge this gap, we introduce BoRP (Bootstrapped Regression Probing), a scalable framework for high-fidelity satisfaction evaluation. Unlike generative approaches, BoRP leverages the geometric properties of LLM latent space. It employs a polarization-index-based bootstrapping mechanism to automate rubric generation and utilizes Partial Least Squares (PLS) to map hidden states to continuous scores. Experiments on industrial datasets show that BoRP (Qwen3-8B/14B) significantly outperforms generative baselines (even Qwen3-Max) in alignment with human judgments. Furthermore, BoRP reduces inference costs by orders of magnitude, enabling full-scale monitoring and highly sensitive A/B testing via CUPED.
title BoRP: Bootstrapped Regression Probing for Scalable and Human-Aligned LLM Evaluation
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2601.18253