Metraipolis › archive › dc_7ce1de2df012
Verifying the Linear Representation Hypothesis: How Interpretable Are Vision SAEs? — arXiv:2609.35020
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.35020
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T23:52:35.774Z
- SHA-256 of the archived text
6208db31b9847faf39fd2305b6fa728fe985a8c03e9dc94ecc5bea05401922ee- Stable URL of this page
https://metraipolis.com/documento/dc_7ce1de2df012
How to cite
Verifying the Linear Representation Hypothesis: How Interpretable Are Vision SAEs? — arXiv:2609.35020 — arXiv (en), https://arxiv.org/abs/2609.35020 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7ce1de2df012 (Metraipolis archive dc_7ce1de2df012, sha256 6208db31b9847faf39fd2305b6fa728fe985a8c03e9dc94ecc5bea05401922ee)
BibTeX
@misc{dc7ce1de2df012,
title = {Verifying the Linear Representation Hypothesis: How Interpretable Are Vision SAEs? — arXiv:2609.35020},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.35020},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7ce1de2df012; sha256 6208db31b9847faf39fd2305b6fa728fe985a8c03e9dc94ecc5bea05401922ee},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35020. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Vision Sparse Autoencoders (SAEs) have become a popular tool in Mechanistic Interpretability due to their presumed ability to disentangle complex features learned by a model into monosemantic concepts. Despite their growing popularity, evaluating their interpretability remains an active topic of research. The bedrock motivating the adoption of SAEs is the Linear Representation Hypothesis (LRH), which claims that polysemantic features can be projected onto a (near) orthogonal basis of sparse, human-understandable representations. Yet, most current frameworks evaluate proxies such as the sparsity of SAE features or the coherence of the inferred dictionary, implicitly assuming that these reflect alignment with human perception. In this paper, we provide empirical evidence that measuring the interpretability of SAE concepts is more difficult than these proxies suggest. To this end, we adapt the Autointerpretability Score (AIS) - previously shown to align with human judgments in Natural Language Processing - to vision tasks and validate our approach in a dedicated user study. We evaluate SAE concept quality using both standard metrics and our adapted AIS. We find that established interpretability metrics for SAEs correlate neither with one another nor with AIS, indicating that no single reference-free metric, whether grounded in the LRH or not, is sufficient for verifying the interpretability of vision SAEs. We argue these findings support recent calls for more verifiable, ground-truth-anchored design and evaluation of explanation methods. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.35020 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T23:29:27.511Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).