{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/from-what-to-how-attributing-clip-s-latent","title":"From What to How: Attributing CLIP's Latent Components Reveals Unexpected Semantic Reliance","arxiv_id":"2505.20229","date":"2025-05-26","proceeding":null,"authors":["Maximilian Dreyer","Lorenz Hufe","Jim Berend","Thomas Wiegand","Sebastian Lapuschkin","Wojciech Samek"],"abstract":"Transformer-based CLIP models are widely used for text-image probing and feature extraction, making it relevant to understand the internal mechanisms behind their predictions. While recent works show that Sparse Autoencoders (SAEs) yield interpretable latent components, they focus on what these encode and miss how they drive predictions. We introduce a scalable framework that reveals what latent components activate for, how they align with expected semantics, and how important they are to predictions. To achieve this, we adapt attribution patching for instance-wise component attributions in CLIP and highlight key faithfulness limitations of the widely used Logit Lens technique. By combining attributions with semantic alignment scores, we can automatically uncover reliance on components that encode semantically unexpected or spurious concepts. Applied across multiple CLIP variants, our method uncovers hundreds of surprising components linked to polysemous words, compound nouns, visual typography and dataset artifacts. While text embeddings remain prone to semantic ambiguity, they are more robust to spurious correlations compared to linear classifiers trained on image embeddings. A case study on skin lesion detection highlights how such classifiers can amplify hidden shortcuts, underscoring the need for holistic, mechanistic interpretability. We provide code at https://github.com/maxdreyer/attributing-clip.","url_abs":"https://arxiv.org/abs/2505.20229v1","url_pdf":"https://arxiv.org/pdf/2505.20229v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"from-what-to-how-attributing-clip-s-latent","repo_url":"https://github.com/maxdreyer/attributing-clip","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"lesion-detection","task_name":"Lesion Detection"}],"methods":[{"method_slug":"align","method_name":"ALIGN"},{"method_slug":"clip","method_name":"CLIP"},{"method_slug":"focus","method_name":"Focus"},{"method_slug":"patching","method_name":"Patching"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2505.20229","atlas_url":"https://app.syntology.ai/?focus=2505.20229","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2505.20229"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2505.20229\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2505.20229"},"call":"get_concepts_for_paper(arxiv_id=\"2505.20229\")"}],"by_repo":[{"repo":"maxdreyer/attributing-clip","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"396c9c1b15fc6ff4e4dc7f460f17524b321fa90e8f6c6573b5ce255823badf15","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}