Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance
Research shows that autointerpretability scores for sparse autoencoders are dominated by evaluation pipeline choices rather than architectural differences, undermining cross-paper comparisons.
-->
[Submitted on 1 Jul 2026]
Title:Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance
View a PDF of the paper titled Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance, by Sinie van der Ben and 3 other authors
View PDF HTML (experimental)
Abstract:Cross-paper comparison of sparse autoencoder (SAE) interpretability often relies on autointerpretability scores. In this evaluation pipeline, a language model (LM) explains each feature, and another LM scores the explanation. For these comparisons to be meaningful, scores must reflect stable properties of the features rather than confounding aspects of the evaluation pipeline. Through systematic experiments across four metrics (simulation, detection, fuzzing, purity), two models (Pythia-160M, Apertus-8B), and four axes of methodological variation, we show that this assumption does not hold. Specifically, we find that R1) methodological variance collectively exceeds architectural variance across all metrics and tested models; R2) each metric exhibits a distinct instability profile, with detection being the most stable and fuzzing unreliable across all conditions; R3) top-k feature rankings do not stay consistent across corpus and draw conditions, masking per-feature instability behind stable mean scores; a failure that cannot be detected by monitoring explanation similarity alone. These findings suggest that cross-paper comparisons based on autointerpretability scores may reflect pipeline differences rather than architectural differences, with implications for the ongoing debate on SAE utility. More broadly, unreliable evaluation slows progress in interpretability research at a time when reliable tools for understanding AI systems are needed. To support evaluation, we contribute a variance decomposition approach, a Stability Check, and a Minimum Reporting Checklist.
Subjects:
Machine Learning (cs.LG); Computation and Language (cs.CL)
Cite as: arXiv:2607.19386 [cs.LG]
(or arXiv:2607.19386v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.19386
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Sinie Van Der Ben [view email] [v1] Wed, 1 Jul 2026 13:18:12 UTC (2,658 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance, by Sinie van der Ben and 3 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.LG
new | recent | 2026-07
Change to browse by:
cs cs.CL
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
Loading...
Data provided by:
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
IArxiv recommender toggle
IArxiv Recommender (What is IArxiv?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)