{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CZM6D4HE7WKWU2W3SHIIKCXSS6","short_pith_number":"pith:CZM6D4HE","schema_version":"1.0","canonical_sha256":"1659e1f0e4fd956a6adb91d0850af297b2ee175c4a0465c9585ed4a37cdec139","source":{"kind":"arxiv","id":"2502.13886","version":1},"attestation_state":"computed","paper":{"title":"Refining embeddings with fill-tuning: data-efficient generalised performance improvements for materials foundation models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CE"],"primary_cat":"cs.LG","authors_text":"Edward O. Pyzer-Knapp, Luke Dicks, Matthew P. Wilson, Nicolas Galichet","submitted_at":"2025-02-19T17:17:13Z","abstract_excerpt":"Pretrained foundation models learn embeddings that can be used for a wide range of downstream tasks. These embeddings optimise general performance, and if insufficiently accurate at a specific task the model can be fine-tuned to improve performance. For all current methodologies this operation necessarily degrades performance on all out-of-distribution tasks. In this work we present 'fill-tuning', a novel methodology to generate datasets for continued pretraining of foundation models that are not suited to a particular downstream task, but instead aim to correct poor regions of the embedding. "},"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":"2502.13886","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-19T17:17:13Z","cross_cats_sorted":["cs.CE"],"title_canon_sha256":"55febc7d62d93010cc0395ae1ec1d01d35b6556eff157b52afc9e1cd9cd6a92c","abstract_canon_sha256":"7ef7f0857683315a20578b136b71bf0296101b057d1c608200b75622d59bee92"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:17:02.033686Z","signature_b64":"MkSWKxIpxDQcXlX1pVNnhFa95sVpP9JflxukTzgR5R31Wk+lz4kTsQTYCW18Tq9XR6fswBJXho3AuuNFppCiBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1659e1f0e4fd956a6adb91d0850af297b2ee175c4a0465c9585ed4a37cdec139","last_reissued_at":"2026-07-05T10:17:02.033333Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:17:02.033333Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Refining embeddings with fill-tuning: data-efficient generalised performance improvements for materials foundation models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CE"],"primary_cat":"cs.LG","authors_text":"Edward O. Pyzer-Knapp, Luke Dicks, Matthew P. Wilson, Nicolas Galichet","submitted_at":"2025-02-19T17:17:13Z","abstract_excerpt":"Pretrained foundation models learn embeddings that can be used for a wide range of downstream tasks. These embeddings optimise general performance, and if insufficiently accurate at a specific task the model can be fine-tuned to improve performance. For all current methodologies this operation necessarily degrades performance on all out-of-distribution tasks. In this work we present 'fill-tuning', a novel methodology to generate datasets for continued pretraining of foundation models that are not suited to a particular downstream task, but instead aim to correct poor regions of the embedding. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13886","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/2502.13886/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":"2502.13886","created_at":"2026-07-05T10:17:02.033392+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.13886v1","created_at":"2026-07-05T10:17:02.033392+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13886","created_at":"2026-07-05T10:17:02.033392+00:00"},{"alias_kind":"pith_short_12","alias_value":"CZM6D4HE7WKW","created_at":"2026-07-05T10:17:02.033392+00:00"},{"alias_kind":"pith_short_16","alias_value":"CZM6D4HE7WKWU2W3","created_at":"2026-07-05T10:17:02.033392+00:00"},{"alias_kind":"pith_short_8","alias_value":"CZM6D4HE","created_at":"2026-07-05T10:17:02.033392+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6","json":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6.json","graph_json":"https://pith.science/api/pith-number/CZM6D4HE7WKWU2W3SHIIKCXSS6/graph.json","events_json":"https://pith.science/api/pith-number/CZM6D4HE7WKWU2W3SHIIKCXSS6/events.json","paper":"https://pith.science/paper/CZM6D4HE"},"agent_actions":{"view_html":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6","download_json":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6.json","view_paper":"https://pith.science/paper/CZM6D4HE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.13886&json=true","fetch_graph":"https://pith.science/api/pith-number/CZM6D4HE7WKWU2W3SHIIKCXSS6/graph.json","fetch_events":"https://pith.science/api/pith-number/CZM6D4HE7WKWU2W3SHIIKCXSS6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6/action/storage_attestation","attest_author":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6/action/author_attestation","sign_citation":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6/action/citation_signature","submit_replication":"https://pith.science/pith/CZM6D4HE7WKWU2W3SHIIKCXSS6/action/replication_record"}},"created_at":"2026-07-05T10:17:02.033392+00:00","updated_at":"2026-07-05T10:17:02.033392+00:00"}