{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:OIAU63CAH6QCKPOMWJ6THLXEWU","short_pith_number":"pith:OIAU63CA","schema_version":"1.0","canonical_sha256":"72014f6c403fa0253dccb27d33aee4b517625c1ceeec3697ffa78025253194f2","source":{"kind":"arxiv","id":"1905.00414","version":4},"attestation_state":"computed","paper":{"title":"Similarity of Neural Network Representations Revisited","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["q-bio.NC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Geoffrey Hinton, Honglak Lee, Mohammad Norouzi, Simon Kornblith","submitted_at":"2019-05-01T17:57:26Z","abstract_excerpt":"Recent work has sought to understand the behavior of neural networks by comparing representations between layers and between different trained models. We examine methods for comparing neural network representations based on canonical correlation analysis (CCA). We show that CCA belongs to a family of statistics for measuring multivariate similarity, but that neither CCA nor any other statistic that is invariant to invertible linear transformation can measure meaningful similarities between representations of higher dimension than the number of data points. We introduce a similarity index that "},"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":"1905.00414","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-01T17:57:26Z","cross_cats_sorted":["q-bio.NC","stat.ML"],"title_canon_sha256":"d75a1b382805ecec7568954e14b151556d936eac3dd4d51359e17f7dc0f15033","abstract_canon_sha256":"f5f2f5a3ca794da7eb3fa7464670f8904133770f294cb40b2182aff6bd406e7f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:40:10.868126Z","signature_b64":"k6//dfNyvJUrRiIzpqkErnC1bgrIFJKdzlptUhAV91TNpcr5Wy8mUTi/vbxCRfNDletpypXmgC3vzh7qr+ZxBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"72014f6c403fa0253dccb27d33aee4b517625c1ceeec3697ffa78025253194f2","last_reissued_at":"2026-05-17T23:40:10.867494Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:40:10.867494Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Similarity of Neural Network Representations Revisited","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["q-bio.NC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Geoffrey Hinton, Honglak Lee, Mohammad Norouzi, Simon Kornblith","submitted_at":"2019-05-01T17:57:26Z","abstract_excerpt":"Recent work has sought to understand the behavior of neural networks by comparing representations between layers and between different trained models. We examine methods for comparing neural network representations based on canonical correlation analysis (CCA). We show that CCA belongs to a family of statistics for measuring multivariate similarity, but that neither CCA nor any other statistic that is invariant to invertible linear transformation can measure meaningful similarities between representations of higher dimension than the number of data points. We introduce a similarity index that "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.00414","kind":"arxiv","version":4},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1905.00414","created_at":"2026-05-17T23:40:10.867593+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.00414v4","created_at":"2026-05-17T23:40:10.867593+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.00414","created_at":"2026-05-17T23:40:10.867593+00:00"},{"alias_kind":"pith_short_12","alias_value":"OIAU63CAH6QC","created_at":"2026-05-18T12:33:24.271573+00:00"},{"alias_kind":"pith_short_16","alias_value":"OIAU63CAH6QCKPOM","created_at":"2026-05-18T12:33:24.271573+00:00"},{"alias_kind":"pith_short_8","alias_value":"OIAU63CA","created_at":"2026-05-18T12:33:24.271573+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":15,"internal_anchor_count":8,"sample":[{"citing_arxiv_id":"2605.23778","citing_title":"The physics of AI weather models","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2412.10665","citing_title":"Pretrained Event Classification Model for High Energy Physics Analysis","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2505.17101","citing_title":"A quantitative analysis of semantic information in deep representations of text and images","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.17704","citing_title":"Toy Combinatorial Interpretability Models Reveal Lottery Tickets in Early Feature Space","ref_index":67,"is_internal_anchor":true},{"citing_arxiv_id":"2605.15306","citing_title":"How Data Augmentation Shapes Neural Representations","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2602.14828","citing_title":"Exploring the limits of pre-trained embeddings in machine-guided protein design: a case study on predicting AAV vector viability","ref_index":62,"is_internal_anchor":true},{"citing_arxiv_id":"2605.15183","citing_title":"When Are Two Networks the Same? Tensor Similarity for Mechanistic Interpretability","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2604.13082","citing_title":"The Long Delay to Arithmetic Generalization: When Learned Representations Outrun Behavior","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11530","citing_title":"Multi-Narrow Transformation as a Single-Model Ensemble: Boundary Conditions, Mechanisms, and Failure Modes","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09949","citing_title":"From Syntax to Semantics: Unveiling the Emergence of Chirality in SMILES Translation Models","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08564","citing_title":"Biological Plausibility and Representational Alignment of Feedback Alignment in Convolutional Networks","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24374","citing_title":"MIPIC: Matryoshka Representation Learning via Self-Distilled Intra-Relational and Progressive Information Chaining","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2209.11895","citing_title":"In-context Learning and Induction Heads","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17663","citing_title":"ATLAS: Constitution-Conditioned Latent Geometry and Redistribution Across Language Models and Neural Perturbation Data","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05907","citing_title":"Decoding Alignment without Encoding Alignment: A critique of similarity analysis in neuroscience","ref_index":30,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU","json":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU.json","graph_json":"https://pith.science/api/pith-number/OIAU63CAH6QCKPOMWJ6THLXEWU/graph.json","events_json":"https://pith.science/api/pith-number/OIAU63CAH6QCKPOMWJ6THLXEWU/events.json","paper":"https://pith.science/paper/OIAU63CA"},"agent_actions":{"view_html":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU","download_json":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU.json","view_paper":"https://pith.science/paper/OIAU63CA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.00414&json=true","fetch_graph":"https://pith.science/api/pith-number/OIAU63CAH6QCKPOMWJ6THLXEWU/graph.json","fetch_events":"https://pith.science/api/pith-number/OIAU63CAH6QCKPOMWJ6THLXEWU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU/action/storage_attestation","attest_author":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU/action/author_attestation","sign_citation":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU/action/citation_signature","submit_replication":"https://pith.science/pith/OIAU63CAH6QCKPOMWJ6THLXEWU/action/replication_record"}},"created_at":"2026-05-17T23:40:10.867593+00:00","updated_at":"2026-05-17T23:40:10.867593+00:00"}