@inproceedings{oza-etal-2026-safer,
title = "{SAFER}: Spectral Correlation Prior-based Attention Signature Flow for Enhanced Robustness in Harmful Prompt Detection",
author = "Oza, Neeva and
Kumar, Kolavennu Mani Kanta Krishna and
Mishra, Rahul",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2194/",
pages = "49877--49904",
ISBN = "979-8-89176-459-0",
abstract = "Most existing harmful prompt detection methods rely on surface-level semantic representations, making them vulnerable to paraphrasing, prone to false positives, and limited in their ability to generalize to novel or adversarial prompts. Recent work has explored the use of model-internal signals, such as gradients or spectral features, to train classifiers that capture latent malicious behavior beyond surface semantics. However, these approaches primarily operate on local or isolated internal features and do not model the global structure of model behavior, which limits their robustness against subtle adversarial manipulations. In this work, we focus on robust harmful prompt detection under adversarial attacks. We propose SAFER, a simple yet effective method that models cross-layer attention signature flow by learning spectral correlation priors between attention heads and constructing a cross-layer attention-head spectral graph. This facilitates explicit modeling of structured attention propagation across layers, allowing the capture of both local and global attention dynamics and leading to improved robustness. SAFER achieves superior performance across multiple benchmark datasets for harmful prompt detection, while demonstrating significantly greater resilience to diverse adversarial attacks compared to prior approaches."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="oza-etal-2026-safer">
<titleInfo>
<title>SAFER: Spectral Correlation Prior-based Attention Signature Flow for Enhanced Robustness in Harmful Prompt Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Neeva</namePart>
<namePart type="family">Oza</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kolavennu</namePart>
<namePart type="given">Mani</namePart>
<namePart type="given">Kanta</namePart>
<namePart type="given">Krishna</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rahul</namePart>
<namePart type="family">Mishra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Most existing harmful prompt detection methods rely on surface-level semantic representations, making them vulnerable to paraphrasing, prone to false positives, and limited in their ability to generalize to novel or adversarial prompts. Recent work has explored the use of model-internal signals, such as gradients or spectral features, to train classifiers that capture latent malicious behavior beyond surface semantics. However, these approaches primarily operate on local or isolated internal features and do not model the global structure of model behavior, which limits their robustness against subtle adversarial manipulations. In this work, we focus on robust harmful prompt detection under adversarial attacks. We propose SAFER, a simple yet effective method that models cross-layer attention signature flow by learning spectral correlation priors between attention heads and constructing a cross-layer attention-head spectral graph. This facilitates explicit modeling of structured attention propagation across layers, allowing the capture of both local and global attention dynamics and leading to improved robustness. SAFER achieves superior performance across multiple benchmark datasets for harmful prompt detection, while demonstrating significantly greater resilience to diverse adversarial attacks compared to prior approaches.</abstract>
<identifier type="citekey">oza-etal-2026-safer</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2194/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>49877</start>
<end>49904</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SAFER: Spectral Correlation Prior-based Attention Signature Flow for Enhanced Robustness in Harmful Prompt Detection
%A Oza, Neeva
%A Kumar, Kolavennu Mani Kanta Krishna
%A Mishra, Rahul
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F oza-etal-2026-safer
%X Most existing harmful prompt detection methods rely on surface-level semantic representations, making them vulnerable to paraphrasing, prone to false positives, and limited in their ability to generalize to novel or adversarial prompts. Recent work has explored the use of model-internal signals, such as gradients or spectral features, to train classifiers that capture latent malicious behavior beyond surface semantics. However, these approaches primarily operate on local or isolated internal features and do not model the global structure of model behavior, which limits their robustness against subtle adversarial manipulations. In this work, we focus on robust harmful prompt detection under adversarial attacks. We propose SAFER, a simple yet effective method that models cross-layer attention signature flow by learning spectral correlation priors between attention heads and constructing a cross-layer attention-head spectral graph. This facilitates explicit modeling of structured attention propagation across layers, allowing the capture of both local and global attention dynamics and leading to improved robustness. SAFER achieves superior performance across multiple benchmark datasets for harmful prompt detection, while demonstrating significantly greater resilience to diverse adversarial attacks compared to prior approaches.
%U https://aclanthology.org/2026.emnlp-main.2194/
%P 49877-49904
Markdown (Informal)
[SAFER: Spectral Correlation Prior-based Attention Signature Flow for Enhanced Robustness in Harmful Prompt Detection](https://aclanthology.org/2026.emnlp-main.2194/) (Oza et al., EMNLP 2026)
ACL