AutoMetrics: Approximate Human Judgements with Automatically Generated Evaluators

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ryan, Michael J., Zhang, Yanzhe, Salunkhe, Amol, Chu, Yi, Xu, Di, Yang, Diyi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917156894015488
author Ryan, Michael J.
Zhang, Yanzhe
Salunkhe, Amol
Chu, Yi
Xu, Di
Yang, Diyi
author_facet Ryan, Michael J.
Zhang, Yanzhe
Salunkhe, Amol
Chu, Yi
Xu, Di
Yang, Diyi
contents Evaluating user-facing AI applications remains a central challenge, especially in open-ended domains such as travel planning, clinical note generation, or dialogue. The gold standard is user feedback (e.g., thumbs up/down) or behavioral signals (e.g., retention), but these are often scarce in prototypes and research projects, or too-slow to use for system optimization. We present AutoMetrics, a framework for synthesizing evaluation metrics under low-data constraints. AutoMetrics combines retrieval from MetricBank, a collection of 48 metrics we curate, with automatically generated LLM-as-a-Judge criteria informed by lightweight human feedback. These metrics are composed via regression to maximize correlation with human signal. AutoMetrics takes you from expensive measures to interpretable automatic metrics. Across 5 diverse tasks, AutoMetrics improves Kendall correlation with human ratings by up to 33.4% over LLM-as-a-Judge while requiring fewer than 100 feedback points. We show that AutoMetrics can be used as a proxy reward to equal effect as a verifiable reward. We release the full AutoMetrics toolkit and MetricBank to accelerate adaptive evaluation of LLM applications.
format Preprint
id arxiv_https___arxiv_org_abs_2512_17267
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AutoMetrics: Approximate Human Judgements with Automatically Generated Evaluators
Ryan, Michael J.
Zhang, Yanzhe
Salunkhe, Amol
Chu, Yi
Xu, Di
Yang, Diyi
Computation and Language
Artificial Intelligence
Evaluating user-facing AI applications remains a central challenge, especially in open-ended domains such as travel planning, clinical note generation, or dialogue. The gold standard is user feedback (e.g., thumbs up/down) or behavioral signals (e.g., retention), but these are often scarce in prototypes and research projects, or too-slow to use for system optimization. We present AutoMetrics, a framework for synthesizing evaluation metrics under low-data constraints. AutoMetrics combines retrieval from MetricBank, a collection of 48 metrics we curate, with automatically generated LLM-as-a-Judge criteria informed by lightweight human feedback. These metrics are composed via regression to maximize correlation with human signal. AutoMetrics takes you from expensive measures to interpretable automatic metrics. Across 5 diverse tasks, AutoMetrics improves Kendall correlation with human ratings by up to 33.4% over LLM-as-a-Judge while requiring fewer than 100 feedback points. We show that AutoMetrics can be used as a proxy reward to equal effect as a verifiable reward. We release the full AutoMetrics toolkit and MetricBank to accelerate adaptive evaluation of LLM applications.
title AutoMetrics: Approximate Human Judgements with Automatically Generated Evaluators
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2512.17267