@inproceedings{puri-etal-2025-fade,
title = "{FADE}: Why Bad Descriptions Happen to Good Features",
author = "Puri, Bruno and
Jain, Aakriti and
Golimblevskaia, Elena and
Kahardipraja, Patrick and
Wiegand, Thomas and
Samek, Wojciech and
Lapuschkin, Sebastian",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.881/",
doi = "10.18653/v1/2025.findings-acl.881",
pages = "17138--17160",
ISBN = "979-8-89176-256-5",
abstract = "Recent advances in mechanistic interpretability have highlighted the potential of automating interpretability pipelines in analyzing the latent representations within LLMs. While this may enhance our understanding of internal mechanisms, the field lacks standardized evaluation methods for assessing the validity of discovered features. We attempt to bridge this gap by introducing **FADE**: Feature Alignment to Description Evaluation, a scalable model-agnostic framework for automatically evaluating feature-to-description alignment. **FADE** evaluates alignment across four key metrics {--} *Clarity, Responsiveness, Purity, and Faithfulness* {--} and systematically quantifies the causes of the misalignment between features and their descriptions. We apply **FADE** to analyze existing open-source feature descriptions and assess key components of automated interpretability pipelines, aiming to enhance the quality of descriptions. Our findings highlight fundamental challenges in generating feature descriptions, particularly for SAEs compared to MLP neurons, providing insights into the limitations and future directions of automated interpretability. We release **FADE** as an open-source package at: [github.com/brunibrun/FADE](https://github.com/brunibrun/FADE)."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="puri-etal-2025-fade">
<titleInfo>
<title>FADE: Why Bad Descriptions Happen to Good Features</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bruno</namePart>
<namePart type="family">Puri</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aakriti</namePart>
<namePart type="family">Jain</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Elena</namePart>
<namePart type="family">Golimblevskaia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Patrick</namePart>
<namePart type="family">Kahardipraja</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thomas</namePart>
<namePart type="family">Wiegand</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wojciech</namePart>
<namePart type="family">Samek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sebastian</namePart>
<namePart type="family">Lapuschkin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>Recent advances in mechanistic interpretability have highlighted the potential of automating interpretability pipelines in analyzing the latent representations within LLMs. While this may enhance our understanding of internal mechanisms, the field lacks standardized evaluation methods for assessing the validity of discovered features. We attempt to bridge this gap by introducing **FADE**: Feature Alignment to Description Evaluation, a scalable model-agnostic framework for automatically evaluating feature-to-description alignment. **FADE** evaluates alignment across four key metrics – *Clarity, Responsiveness, Purity, and Faithfulness* – and systematically quantifies the causes of the misalignment between features and their descriptions. We apply **FADE** to analyze existing open-source feature descriptions and assess key components of automated interpretability pipelines, aiming to enhance the quality of descriptions. Our findings highlight fundamental challenges in generating feature descriptions, particularly for SAEs compared to MLP neurons, providing insights into the limitations and future directions of automated interpretability. We release **FADE** as an open-source package at: [github.com/brunibrun/FADE](https://github.com/brunibrun/FADE).</abstract>
<identifier type="citekey">puri-etal-2025-fade</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.881</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.881/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>17138</start>
<end>17160</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T FADE: Why Bad Descriptions Happen to Good Features
%A Puri, Bruno
%A Jain, Aakriti
%A Golimblevskaia, Elena
%A Kahardipraja, Patrick
%A Wiegand, Thomas
%A Samek, Wojciech
%A Lapuschkin, Sebastian
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F puri-etal-2025-fade
%X Recent advances in mechanistic interpretability have highlighted the potential of automating interpretability pipelines in analyzing the latent representations within LLMs. While this may enhance our understanding of internal mechanisms, the field lacks standardized evaluation methods for assessing the validity of discovered features. We attempt to bridge this gap by introducing **FADE**: Feature Alignment to Description Evaluation, a scalable model-agnostic framework for automatically evaluating feature-to-description alignment. **FADE** evaluates alignment across four key metrics – *Clarity, Responsiveness, Purity, and Faithfulness* – and systematically quantifies the causes of the misalignment between features and their descriptions. We apply **FADE** to analyze existing open-source feature descriptions and assess key components of automated interpretability pipelines, aiming to enhance the quality of descriptions. Our findings highlight fundamental challenges in generating feature descriptions, particularly for SAEs compared to MLP neurons, providing insights into the limitations and future directions of automated interpretability. We release **FADE** as an open-source package at: [github.com/brunibrun/FADE](https://github.com/brunibrun/FADE).
%R 10.18653/v1/2025.findings-acl.881
%U https://aclanthology.org/2025.findings-acl.881/
%U https://doi.org/10.18653/v1/2025.findings-acl.881
%P 17138-17160
Markdown (Informal)
[FADE: Why Bad Descriptions Happen to Good Features](https://aclanthology.org/2025.findings-acl.881/) (Puri et al., Findings 2025)
ACL
- Bruno Puri, Aakriti Jain, Elena Golimblevskaia, Patrick Kahardipraja, Thomas Wiegand, Wojciech Samek, and Sebastian Lapuschkin. 2025. FADE: Why Bad Descriptions Happen to Good Features. In Findings of the Association for Computational Linguistics: ACL 2025, pages 17138–17160, Vienna, Austria. Association for Computational Linguistics.