{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:DG4D6K74C4NWJKHGD2GMAXS7L5","short_pith_number":"pith:DG4D6K74","schema_version":"1.0","canonical_sha256":"19b83f2bfc171b64a8e61e8cc05e5f5f665524dfc1a9cd1aa1a0d26e5b7b07c9","source":{"kind":"arxiv","id":"2106.07682","version":1},"attestation_state":"computed","paper":{"title":"Revisiting Model Stitching to Compare Neural Representations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Boaz Barak, Preetum Nakkiran, Yamini Bansal","submitted_at":"2021-06-14T18:05:10Z","abstract_excerpt":"We revisit and extend model stitching (Lenc & Vedaldi 2015) as a methodology to study the internal representations of neural networks. Given two trained and frozen models $A$ and $B$, we consider a \"stitched model'' formed by connecting the bottom-layers of $A$ to the top-layers of $B$, with a simple trainable layer between them. We argue that model stitching is a powerful and perhaps under-appreciated tool, which reveals aspects of representations that measures such as centered kernel alignment (CKA) cannot. Through extensive experiments, we use model stitching to obtain quantitative verifica"},"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":"2106.07682","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-14T18:05:10Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"54f2b6b12dab353fcc2592b390286fe27a557f6bee1bb0d746a5b68d3e069a87","abstract_canon_sha256":"e30bf03f182150b0e18c49ed2889dd50d3264acace33bd0c298be737451767b8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:49:32.738781Z","signature_b64":"UBF8oi7Gfg6dKsD00KfcQZHATPzRMyV+cnndxUm7AulGdQKJIcBxzZFYxhS3K1l2xeMV7zHVFoaHELu/0JNiBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"19b83f2bfc171b64a8e61e8cc05e5f5f665524dfc1a9cd1aa1a0d26e5b7b07c9","last_reissued_at":"2026-07-05T02:49:32.738368Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:49:32.738368Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Revisiting Model Stitching to Compare Neural Representations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Boaz Barak, Preetum Nakkiran, Yamini Bansal","submitted_at":"2021-06-14T18:05:10Z","abstract_excerpt":"We revisit and extend model stitching (Lenc & Vedaldi 2015) as a methodology to study the internal representations of neural networks. Given two trained and frozen models $A$ and $B$, we consider a \"stitched model'' formed by connecting the bottom-layers of $A$ to the top-layers of $B$, with a simple trainable layer between them. We argue that model stitching is a powerful and perhaps under-appreciated tool, which reveals aspects of representations that measures such as centered kernel alignment (CKA) cannot. Through extensive experiments, we use model stitching to obtain quantitative verifica"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.07682","kind":"arxiv","version":1},"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/2106.07682/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":"2106.07682","created_at":"2026-07-05T02:49:32.738432+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.07682v1","created_at":"2026-07-05T02:49:32.738432+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.07682","created_at":"2026-07-05T02:49:32.738432+00:00"},{"alias_kind":"pith_short_12","alias_value":"DG4D6K74C4NW","created_at":"2026-07-05T02:49:32.738432+00:00"},{"alias_kind":"pith_short_16","alias_value":"DG4D6K74C4NWJKHG","created_at":"2026-07-05T02:49:32.738432+00:00"},{"alias_kind":"pith_short_8","alias_value":"DG4D6K74","created_at":"2026-07-05T02:49:32.738432+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22019","citing_title":"Channel Location Constrains the Auditability of Subliminal Learning","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03280","citing_title":"A Negative Result on Cross-Model Activation Transfer in a Pythia Multi-Hop Setting","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25225","citing_title":"Transformer Field Theory: A Response-Theoretic Approach to Mechanistic Interpretability","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08335","citing_title":"Dead Weights, Live Signals: Feedforward Graphs of Frozen Language Models","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06017","citing_title":"Toward Aristotelian Medical Representations: Backpropagation-Free Layer-wise Analysis for Interpretable Generalized Metric Learning on MedMNIST","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01048","citing_title":"Compared to What? Baselines and Metrics for Counterfactual Prompting","ref_index":78,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5","json":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5.json","graph_json":"https://pith.science/api/pith-number/DG4D6K74C4NWJKHGD2GMAXS7L5/graph.json","events_json":"https://pith.science/api/pith-number/DG4D6K74C4NWJKHGD2GMAXS7L5/events.json","paper":"https://pith.science/paper/DG4D6K74"},"agent_actions":{"view_html":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5","download_json":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5.json","view_paper":"https://pith.science/paper/DG4D6K74","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.07682&json=true","fetch_graph":"https://pith.science/api/pith-number/DG4D6K74C4NWJKHGD2GMAXS7L5/graph.json","fetch_events":"https://pith.science/api/pith-number/DG4D6K74C4NWJKHGD2GMAXS7L5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5/action/storage_attestation","attest_author":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5/action/author_attestation","sign_citation":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5/action/citation_signature","submit_replication":"https://pith.science/pith/DG4D6K74C4NWJKHGD2GMAXS7L5/action/replication_record"}},"created_at":"2026-07-05T02:49:32.738432+00:00","updated_at":"2026-07-05T02:49:32.738432+00:00"}