{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:6EMEYRV24RFLKZEAURZ7DXNCNZ","short_pith_number":"pith:6EMEYRV2","schema_version":"1.0","canonical_sha256":"f1184c46bae44ab56480a473f1dda26e745acb85f0df4638b6a6671e5ec2e20f","source":{"kind":"arxiv","id":"2108.01661","version":2},"attestation_state":"computed","paper":{"title":"Grounding Representation Similarity with Statistical Testing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Frances Ding, Jacob Steinhardt, Jean-Stanislas Denain","submitted_at":"2021-08-03T17:58:16Z","abstract_excerpt":"To understand neural network behavior, recent works quantitatively compare different networks' learned representations using canonical correlation analysis (CCA), centered kernel alignment (CKA), and other dissimilarity measures. Unfortunately, these widely used measures often disagree on fundamental observations, such as whether deep networks differing only in random initialization learn similar representations. These disagreements raise the question: which, if any, of these dissimilarity measures should we believe? We provide a framework to ground this question through a concrete test: measu"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2108.01661","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-08-03T17:58:16Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"60058295f46a1113da682264716754702b97713994f0e793a5a37fef63dd5ba2","abstract_canon_sha256":"8d6e6ffac97b6ce31f688aee15e553e0e39162859d9ecae6474fc212580d95fd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:28:40.586967Z","signature_b64":"XEMF8QKdIcIbyLDLNmSrOldZIaK0IMZ2okNMgQ6zscEpjIPjqA3pupSXBzXM2q3IjsCm29PwH2JQqoEyEFr5Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f1184c46bae44ab56480a473f1dda26e745acb85f0df4638b6a6671e5ec2e20f","last_reissued_at":"2026-07-05T03:28:40.586449Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:28:40.586449Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Grounding Representation Similarity with Statistical Testing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Frances Ding, Jacob Steinhardt, Jean-Stanislas Denain","submitted_at":"2021-08-03T17:58:16Z","abstract_excerpt":"To understand neural network behavior, recent works quantitatively compare different networks' learned representations using canonical correlation analysis (CCA), centered kernel alignment (CKA), and other dissimilarity measures. Unfortunately, these widely used measures often disagree on fundamental observations, such as whether deep networks differing only in random initialization learn similar representations. These disagreements raise the question: which, if any, of these dissimilarity measures should we believe? We provide a framework to ground this question through a concrete test: measu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2108.01661","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2108.01661/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2108.01661","created_at":"2026-07-05T03:28:40.586536+00:00"},{"alias_kind":"arxiv_version","alias_value":"2108.01661v2","created_at":"2026-07-05T03:28:40.586536+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2108.01661","created_at":"2026-07-05T03:28:40.586536+00:00"},{"alias_kind":"pith_short_12","alias_value":"6EMEYRV24RFL","created_at":"2026-07-05T03:28:40.586536+00:00"},{"alias_kind":"pith_short_16","alias_value":"6EMEYRV24RFLKZEA","created_at":"2026-07-05T03:28:40.586536+00:00"},{"alias_kind":"pith_short_8","alias_value":"6EMEYRV2","created_at":"2026-07-05T03:28:40.586536+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.05907","citing_title":"Decoding Alignment without Encoding Alignment: A critique of similarity analysis in neuroscience","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ","json":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ.json","graph_json":"https://pith.science/api/pith-number/6EMEYRV24RFLKZEAURZ7DXNCNZ/graph.json","events_json":"https://pith.science/api/pith-number/6EMEYRV24RFLKZEAURZ7DXNCNZ/events.json","paper":"https://pith.science/paper/6EMEYRV2"},"agent_actions":{"view_html":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ","download_json":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ.json","view_paper":"https://pith.science/paper/6EMEYRV2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2108.01661&json=true","fetch_graph":"https://pith.science/api/pith-number/6EMEYRV24RFLKZEAURZ7DXNCNZ/graph.json","fetch_events":"https://pith.science/api/pith-number/6EMEYRV24RFLKZEAURZ7DXNCNZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ/action/storage_attestation","attest_author":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ/action/author_attestation","sign_citation":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ/action/citation_signature","submit_replication":"https://pith.science/pith/6EMEYRV24RFLKZEAURZ7DXNCNZ/action/replication_record"}},"created_at":"2026-07-05T03:28:40.586536+00:00","updated_at":"2026-07-05T03:28:40.586536+00:00"}