{"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/arxiv-2606-06333","title":"Subspace-Aware Sparse Autoencoders for Effective Mechanistic Interpretability","arxiv_id":"2606.06333","date":"2026-06-04","proceeding":null,"authors":["Seyed Arshan Dalili","Mehrdad Mahdavi"],"abstract":"Sparse Autoencoders (SAEs) are widely used for mechanistic interpretability in large language models, yet their formulation assigns each latent feature a single decoder direction, implicitly assuming features to be one-dimensional. We show that this assumption mismatches with the multi-dimensional structure of model features, provably inducing feature splitting through two distinct mechanisms. Geometrically, reconstructing a feature of intrinsic dimension $d_i \\ge 2$ to error $\\varepsilon$ with single-direction decoders forces a number of atoms that is exponential in $d_i$. From an end-to-end optimization perspective, this splitting is not merely possible but actively preferred. We prove that there exists a continuous path from the true $d_i$-dimensional basis to a strictly lower risk of the $\\ell_1$-regularized SAE objective, whose descent directions drive any trained dictionary into that exponential regime. A single coherent feature is therefore fragmented across many near-collinear latents, producing spurious multiplicity and obscuring the intrinsic geometry. Motivated by this, we introduce Subspace-Aware Sparse Autoencoders (SASA), which replace single-vector decoders with learned decoder subspaces, enforce block sparsity via Top-$s$ group gating, and adapt each group's effective rank with a nuclear-norm regularizer. We then show that once the block size satisfies $r \\ge d_i$, a single group not only can represent the entire feature slice but is the global minimizer of the SASA objective. This consolidation yields a sample complexity polynomial in $d_i$ rather than exponential -- a decisive advantage given that every training activation costs an LLM forward pass. Empirically, on GPT-2 and Mistral-7B, SASA reduces feature splitting and absorption, improves monosemanticity and interpretability, and matches or exceeds standard SAEs while training on roughly half the token budget.","url_abs":"https://arxiv.org/abs/2606.06333","url_pdf":"https://arxiv.org/pdf/2606.06333","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2606.06333","atlas_url":"https://app.syntology.ai/?focus=2606.06333","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2606.06333"},"calls":[{"tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2606.06333"},"returns":"the samples harvested from this paper, each with its run record; it does not verify the paper"},{"tool":"get_code","arguments_in":"samples[].code_sha256","returns":"one stored sample by its code_sha256; not shown on this site"}]},"developers":"https://syntology.ai/developers","held":true,"reason":"in_graph_layer","checked_against":null,"agent_calls":[{"tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2606.06333"},"call":"get_harvested_code_for_paper(arxiv_id=\"2606.06333\")"},{"tool":"get_code","arguments_in":{"code_sha256":"syntology.samples[].code_sha256"},"call":"get_code(code_sha256=\"…\")"},{"tool":"get_citation_path","arguments":{"paper_1":"2606.06333"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2606.06333\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2606.06333"},"call":"get_concepts_for_paper(arxiv_id=\"2606.06333\")"}],"by_repo":[{"repo":"arshandalili/sasa","kind":"found_in_text","samples":11,"ran":8,"instrument":7,"unverified":3}],"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","read_at":"2026-10-05T00:36:11+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","ran_record":{"url":"https://syntology.ai/api/ran/2606.06333.json","schema":"syntology.ran_record/1","counts_crosswalk":{"n_samples":"syntology.counts.lifted","n_ran":"syntology.counts.executed","n_ran_checked":"syntology.counts.checked","n_instrument":"syntology.counts.instrument_failures","n_unverified":"syntology.counts.no_recorded_run"},"note":"live record, measured at request time; may be newer than this page's graph read","badge":"https://syntology.ai/api/ran/2606.06333.svg"},"claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/arshandalili/sasa","reach":null}],"summary":{"ran":1,"ran_fixture":1,"ran_draft_wrong":6,"unverified":3},"n_samples":11,"n_ran":8,"n_constructed":0,"n_ran_checked":1,"n_instrument":7,"n_unverified":3,"n_instrument_is":"failures of Syntology's instrument, not of the code","every_run_is_an_instrument_failure":false,"key_notes":{"n_ran_checked":"legacy name, kept unchanged so existing readers do not break: it counts the samples that ran with no instrument failure (honoured, violated, and ran with no contract checked); it does not mean a contract was checked, and the pages print it as 'K with no instrument failure', not 'K checked'","n_constructed":"a sub-count of the samples that ran, never subtracted from them and never a failure: an executed sample whose run returned an instance of its own class (fixture_out_type equals the entry name): the run built an object and did not compute a result (Syntology's RAN record, counts.constructed)"},"by_repo_kind":{"found_in_text":{"samples":11,"ran":8,"n_ran_checked":1,"n_instrument":7,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":11,"inline_ok_is":"inline_ok is true only when both licence gates allow Syntology to serve this code's text: this paper's link (inline_ok_link, from the licence of this paper's copy) and the stored copy of the same code that Syntology serves, which may have been first harvested from another repository. withheld_clause is Syntology's own wording of which gate keeps the text a pointer and why, and is null when inline_ok is true. This site shows no code text either way; n_pointer_only_for_licence counts samples with inline_ok false (both gates), n_pointer_only_link_allows those of them whose inline_ok_link is true","n_pointer_only_link_allows":0,"fixture_fields_is":{"signature_and_fixture_out_type":"the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right","level_recorded_at":"when Syntology's graph wrote the sample's verification level, not when the code executed; absent where the level was written before the date was kept, and never filled in with the build date","present_on":"samples that ran only; a key is absent, never empty, where the graph records nothing"},"n_ran_with_level_date":8,"n_ran_without_level_date":0,"n_ran_signature_not_shown":0,"levels_written_between":{"first":"2026-09-24","last":"2026-09-24","is":"the first and last UTC dates on which Syntology's graph wrote a verification level for a sample on this page that ran, not when the code executed"},"code_page_min_papers":5,"code_page_is":"a page for the entry name across every paper's harvested code; the name is a string, not an identity","samples":[{"code_sha256":"2fac6816fcedb9cf","code_sha256_prefix":"2fac6816fcedb9cf","entry":"load_split","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"eval/absorption.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/eval/absorption.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"signature":"probe_dir:PosixPath/0/?, split:str/0/str, tokenizer:MinimalTokenizer/0/?, device:str/0/str","fixture_out_type":"tuple","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","code_page":"/code/load-split","census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"2fac6816fcedb9cf"}},{"code_sha256":"4291a5d025460f82","code_sha256_prefix":"4291a5d025460f82","entry":"pca_dimension","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"analysis/cluster_decoders.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/analysis/cluster_decoders.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":false,"signature":"Xg:ndarray/2/float, var_threshold:float/0/float, center:bool/0/bool","fixture_out_type":"int","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"4291a5d025460f82"}},{"code_sha256":"7e6084a3a5f0c07f","code_sha256_prefix":"7e6084a3a5f0c07f","entry":"read_json","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"common.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/common.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"signature":"path:str/0/str","fixture_out_type":"dict","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","code_page":"/code/read-json","census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"7e6084a3a5f0c07f"}},{"code_sha256":"52ed43d50830b060","code_sha256_prefix":"52ed43d50830b060","entry":"resolve_device","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"common.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/common.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"signature":"device:NoneType/0/?","fixture_out_type":"str","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","code_page":"/code/resolve-device","census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"52ed43d50830b060"}},{"code_sha256":"c3168ff0b414d637","code_sha256_prefix":"c3168ff0b414d637","entry":"single_token_ids","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"analysis/temporal_subspace.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/analysis/temporal_subspace.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"signature":"tokenizer:StandInTokenizer/0/?, terms:list/1/str","fixture_out_type":"dict","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"c3168ff0b414d637"}},{"code_sha256":"d6a09534d473bdf8","code_sha256_prefix":"d6a09534d473bdf8","entry":"slugify_label","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"common.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/common.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":true,"signature":"text:str/0/str","fixture_out_type":"str","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","retrieve_with":{"tool":"get_code","code_sha256":"d6a09534d473bdf8"},"code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"d6a09534d473bdf8"}},{"code_sha256":"0ed6143cc259a5ff","code_sha256_prefix":"0ed6143cc259a5ff","entry":"spectral_metrics_from_values","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"analysis/rank_analysis.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/analysis/rank_analysis.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"signature":"values:ndarray/1/float, energy_threshold:float/0/float","fixture_out_type":"dict","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"0ed6143cc259a5ff"}},{"code_sha256":"dda600f8de06c654","code_sha256_prefix":"dda600f8de06c654","entry":"temporal_vocabulary","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"analysis/temporal_subspace.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/analysis/temporal_subspace.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"signature":"tokenizer:MockTokenizer/0/?, year_min:int/0/int, year_max:int/0/int","fixture_out_type":"dict","level_recorded_at":"2026-09-24T15:16:18.137819+00:00","code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"dda600f8de06c654"}},{"code_sha256":"7ed5392a8f60b7cc","code_sha256_prefix":"7ed5392a8f60b7cc","entry":"group_coords","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"analysis/temporal_subspace.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/analysis/temporal_subspace.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"7ed5392a8f60b7cc"}},{"code_sha256":"c6b764437a26dfad","code_sha256_prefix":"c6b764437a26dfad","entry":"load_ground_truth_probe","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"eval/absorption.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/eval/absorption.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"c6b764437a26dfad"}},{"code_sha256":"d8dc02e5116708d3","code_sha256_prefix":"d8dc02e5116708d3","entry":"resolve_api_key","repo":"arshandalili/sasa","repo_kind":"found_in_text","path":"eval/run_autointerp.py","file_url":"https://github.com/arshandalili/sasa/blob/HEAD/eval/run_autointerp.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"code_page":null,"census":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, arshandalili/sasa when it was checked","mcp_get_code":{"code_sha256":"d8dc02e5116708d3"}}]},"code_links_note":"this paper is newer than the archive, which holds no code links for it; its repositories, from Syntology's graph, are under syntology.repos","arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":{"kind":"leaderboard_placements","source":"Syntology's leaderboard-shaped extractor over the paper's own arXiv-HTML tables: a model pointed at a cell, the number was read from that cell and checked against the board's metric, dataset, split and scale, and an independent check accepted the entry; not reviewed by the paper's authors or the archive's editors","extractor_model":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","verifier_model":null,"prompt_sha":"fa63d4bb9d755694","coverage":{"sentence":"Syntology has checked 8,886 of the 18,637 papers on this site that are newer than the archive (for 3,717 of them no archive leaderboard matched the paper's tables, so there was nothing further to check); 733 were read and have nothing to place (arXiv has no HTML version of the paper, or that version has no tables), 41 could not be read (the extractor's reply could not be parsed), and results from the other 8,977 appear after they are checked.","papers_newer_than_archive":18637,"papers_checked":8886,"papers_read_nothing_to_place":733,"papers_could_not_be_read":41},"entries":[],"judged_to_report_on":{"model":"Claude Sonnet 4.5","model_as_recorded":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","note":"a model's judgement from the paper's own tables, not a result; outcome is what Syntology's checks did with it; reason is the rules' plain words, only for refused_by_rule","boards":[]},"not_placed":{"boards":0,"rejected_by_independent_check":0,"refused_by_a_rule":0,"check_did_not_answer":0,"proposed_without_a_cell":0,"declined_by_site":0}},"record_sha256":"6b246da8c493e13c125b5818b2616e92eb94456d2a31b13b7dc05accd92b306a","record_changed_at":"2026-10-05","record_changed_at_basis":"changed"}