Saved in:
Bibliographic Details
Main Authors: Lin, Zheyu, Yang, Jirui, Qiu, Yukui, Guo, Hengqi, Bao, Yubing, Guan, Yao
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.14195
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912810116579328
author Lin, Zheyu
Yang, Jirui
Qiu, Yukui
Guo, Hengqi
Bao, Yubing
Guan, Yao
author_facet Lin, Zheyu
Yang, Jirui
Qiu, Yukui
Guo, Hengqi
Bao, Yubing
Guan, Yao
contents Evaluating the safety robustness of LLMs is critical for their deployment. However, mainstream Red Teaming methods rely on online generation and black-box output analysis. These approaches are not only costly but also suffer from feedback latency, making them unsuitable for agile diagnostics after training a new model. To address this, we propose N-GLARE (A Non-Generative, Latent Representation-Efficient LLM Safety Evaluator). N-GLARE operates entirely on the model's latent representations, bypassing the need for full text generation. It characterizes hidden layer dynamics by analyzing the APT (Angular-Probabilistic Trajectory) of latent representations and introducing the JSS (Jensen-Shannon Separability) metric. Experiments on over 40 models and 20 red teaming strategies demonstrate that the JSS metric exhibits high consistency with the safety rankings derived from Red Teaming. N-GLARE reproduces the discriminative trends of large-scale red-teaming tests at less than 1\% of the token cost and the runtime cost, providing an efficient output-free evaluation proxy for real-time diagnostics.
format Preprint
id arxiv_https___arxiv_org_abs_2511_14195
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle N-GLARE: An Non-Generative Latent Representation-Efficient LLM Safety Evaluator
Lin, Zheyu
Yang, Jirui
Qiu, Yukui
Guo, Hengqi
Bao, Yubing
Guan, Yao
Machine Learning
Cryptography and Security
Evaluating the safety robustness of LLMs is critical for their deployment. However, mainstream Red Teaming methods rely on online generation and black-box output analysis. These approaches are not only costly but also suffer from feedback latency, making them unsuitable for agile diagnostics after training a new model. To address this, we propose N-GLARE (A Non-Generative, Latent Representation-Efficient LLM Safety Evaluator). N-GLARE operates entirely on the model's latent representations, bypassing the need for full text generation. It characterizes hidden layer dynamics by analyzing the APT (Angular-Probabilistic Trajectory) of latent representations and introducing the JSS (Jensen-Shannon Separability) metric. Experiments on over 40 models and 20 red teaming strategies demonstrate that the JSS metric exhibits high consistency with the safety rankings derived from Red Teaming. N-GLARE reproduces the discriminative trends of large-scale red-teaming tests at less than 1\% of the token cost and the runtime cost, providing an efficient output-free evaluation proxy for real-time diagnostics.
title N-GLARE: An Non-Generative Latent Representation-Efficient LLM Safety Evaluator
topic Machine Learning
Cryptography and Security
url https://arxiv.org/abs/2511.14195