{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:64SV5BSOZBP3Z432G3DIVXP2FH","short_pith_number":"pith:64SV5BSO","schema_version":"1.0","canonical_sha256":"f7255e864ec85fbcf37a36c68addfa29c1200a39eb8de91d656f6b98fc79e1c4","source":{"kind":"arxiv","id":"2512.01766","version":2},"attestation_state":"computed","paper":{"title":"On the Unreasonable Effectiveness of Last-layer Retraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Last-layer retraining succeeds mainly because the held-out set has better group balance than the full training data.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"John C. Hill, Tyler LaBonte, Vidya Muthukumar, Xinchen Zhang","submitted_at":"2025-12-01T15:08:43Z","abstract_excerpt":"Last-layer retraining (LLR) methods -- wherein the last layer of a neural network is reinitialized and retrained on a held-out set following ERM training -- have garnered interest as an efficient approach to rectify dependence on spurious correlations and improve performance on minority groups. Surprisingly, LLR has been found to improve worst-group accuracy even when the held-out set is an imbalanced subset of the training set. We initially hypothesize that this ``unreasonable effectiveness'' of LLR is explained by its ability to mitigate neural collapse through the held-out set, resulting in"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2512.01766","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-12-01T15:08:43Z","cross_cats_sorted":[],"title_canon_sha256":"4c1784338706bb0534f91fa5c9a711ef284fd1913be679bb249fca138a7e1b02","abstract_canon_sha256":"ac75b42556910ed9a5a2a132fcb3231d6691f4b187e09d64cb8ee5f3262283f3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:16.975084Z","signature_b64":"JRqxxpv+Oe4cu0kCHXPsZbvpsvJoKEUBfxQ9Bst2MvykL9NWVemK0haUpjp8lg3KMOMkVmm05Ci9LQO0eUauDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f7255e864ec85fbcf37a36c68addfa29c1200a39eb8de91d656f6b98fc79e1c4","last_reissued_at":"2026-05-17T23:39:16.974390Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:16.974390Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Unreasonable Effectiveness of Last-layer Retraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Last-layer retraining succeeds mainly because the held-out set has better group balance than the full training data.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"John C. Hill, Tyler LaBonte, Vidya Muthukumar, Xinchen Zhang","submitted_at":"2025-12-01T15:08:43Z","abstract_excerpt":"Last-layer retraining (LLR) methods -- wherein the last layer of a neural network is reinitialized and retrained on a held-out set following ERM training -- have garnered interest as an efficient approach to rectify dependence on spurious correlations and improve performance on minority groups. Surprisingly, LLR has been found to improve worst-group accuracy even when the held-out set is an imbalanced subset of the training set. We initially hypothesize that this ``unreasonable effectiveness'' of LLR is explained by its ability to mitigate neural collapse through the held-out set, resulting in"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We present strong evidence for an alternative hypothesis: that the success of LLR is primarily due to better group balance in the held-out set.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the empirical tests on neural collapse are conclusive enough to rule it out as a contributing factor, and that group balance differences are the dominant cause rather than other unmeasured effects.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Last-layer retraining succeeds primarily because the held-out set has better group balance, supported by experiments that do not back the neural collapse hypothesis.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Last-layer retraining succeeds mainly because the held-out set has better group balance than the full training data.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"0b7b3eb0347cea5ea0b1bf0d3f0057551bbffbe3ec6c61c48e1da7a07544aaf7"},"source":{"id":"2512.01766","kind":"arxiv","version":2},"verdict":{"id":"f57ac31a-9834-42de-a042-3529860fa6c4","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T02:39:21.315834Z","strongest_claim":"We present strong evidence for an alternative hypothesis: that the success of LLR is primarily due to better group balance in the held-out set.","one_line_summary":"Last-layer retraining succeeds primarily because the held-out set has better group balance, supported by experiments that do not back the neural collapse hypothesis.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the empirical tests on neural collapse are conclusive enough to rule it out as a contributing factor, and that group balance differences are the dominant cause rather than other unmeasured effects.","pith_extraction_headline":"Last-layer retraining succeeds mainly because the held-out set has better group balance than the full training data."},"references":{"count":4,"sample":[{"doi":"10.1073/pnas.2103091118.url:http://dx.doi.org/10.1073/","year":1998,"title":"Boosting the margin: A new explanation for the effectiveness of voting methods","work_id":"18928c48-673b-4b24-8e73-8e2efe177a6d","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"A broad-coverage challenge corpus for sentence understanding through inference","work_id":"ab1ad3fa-51ee-4b7d-ba96-8f51078051f8","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1992,"title":"Note that Waterbirds is the only dataset that has a distribution shift and MultiNLI is the only dataset which is class-balanceda priori","work_id":"2f29390d-9bd5-42f4-9409-90fa8f21e077","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"These pretrained models are used as the initialization for ERM finetuning under the cross-entropy loss","work_id":"279fb77b-4749-4b2f-93d8-9068c2720b07","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":4,"snapshot_sha256":"cce3bac278234ceef620b01009cf6196a9d54eef76e558c96f07bb8e267df06d","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"02ffadf4474dddc8cb5b9e114b7b7cc721cf2646dcbd60582f5b320aab2fcbb6"},"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":"2512.01766","created_at":"2026-05-17T23:39:16.974513+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.01766v2","created_at":"2026-05-17T23:39:16.974513+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.01766","created_at":"2026-05-17T23:39:16.974513+00:00"},{"alias_kind":"pith_short_12","alias_value":"64SV5BSOZBP3","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"64SV5BSOZBP3Z432","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"64SV5BSO","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH","json":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH.json","graph_json":"https://pith.science/api/pith-number/64SV5BSOZBP3Z432G3DIVXP2FH/graph.json","events_json":"https://pith.science/api/pith-number/64SV5BSOZBP3Z432G3DIVXP2FH/events.json","paper":"https://pith.science/paper/64SV5BSO"},"agent_actions":{"view_html":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH","download_json":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH.json","view_paper":"https://pith.science/paper/64SV5BSO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.01766&json=true","fetch_graph":"https://pith.science/api/pith-number/64SV5BSOZBP3Z432G3DIVXP2FH/graph.json","fetch_events":"https://pith.science/api/pith-number/64SV5BSOZBP3Z432G3DIVXP2FH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH/action/storage_attestation","attest_author":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH/action/author_attestation","sign_citation":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH/action/citation_signature","submit_replication":"https://pith.science/pith/64SV5BSOZBP3Z432G3DIVXP2FH/action/replication_record"}},"created_at":"2026-05-17T23:39:16.974513+00:00","updated_at":"2026-05-17T23:39:16.974513+00:00"}