{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:DM4MZW2ZOGSVYQXVWAB77I4IFI","short_pith_number":"pith:DM4MZW2Z","schema_version":"1.0","canonical_sha256":"1b38ccdb5971a55c42f5b003ffa3882a1b3067af2e975cce0b9cb47d1a95c273","source":{"kind":"arxiv","id":"2211.12092","version":1},"attestation_state":"computed","paper":{"title":"Linear Interpolation In Parameter Space is Good Enough for Fine-Tuned Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Daniil Gavrilov, Mark Rofin, Nikita Balagansky","submitted_at":"2022-11-22T08:49:22Z","abstract_excerpt":"The simplest way to obtain continuous interpolation between two points in high dimensional space is to draw a line between them. While previous works focused on the general connectivity between model parameters, we explored linear interpolation for parameters of pre-trained models after fine-tuning. Surprisingly, we could perform linear interpolation without a performance drop in intermediate points for fine-tuned models. For controllable text generation, such interpolation could be seen as moving a model towards or against the desired text attribute (e.g., positive sentiment), which could be "},"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":"2211.12092","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-11-22T08:49:22Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6e2f97c66179c4bd4a38169a5d93111f5502055d83d2b95869f895f989d9e88c","abstract_canon_sha256":"e8d2ec9a6fdaf1511b50c5749bbc94cc3b8d09b7f5827161cb17b46985f19807"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:18:13.002219Z","signature_b64":"cLCGaBaLjoI8qvZHAH2wr3uV7mEqyPQ/SgT2Onobvn+AIJ4MboGlWC7xyB3YFM2kRVsLc24+EtqSXtVFE66ABw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1b38ccdb5971a55c42f5b003ffa3882a1b3067af2e975cce0b9cb47d1a95c273","last_reissued_at":"2026-07-05T05:18:13.001795Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:18:13.001795Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Linear Interpolation In Parameter Space is Good Enough for Fine-Tuned Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Daniil Gavrilov, Mark Rofin, Nikita Balagansky","submitted_at":"2022-11-22T08:49:22Z","abstract_excerpt":"The simplest way to obtain continuous interpolation between two points in high dimensional space is to draw a line between them. While previous works focused on the general connectivity between model parameters, we explored linear interpolation for parameters of pre-trained models after fine-tuning. Surprisingly, we could perform linear interpolation without a performance drop in intermediate points for fine-tuned models. For controllable text generation, such interpolation could be seen as moving a model towards or against the desired text attribute (e.g., positive sentiment), which could be "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.12092","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/2211.12092/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":"2211.12092","created_at":"2026-07-05T05:18:13.001852+00:00"},{"alias_kind":"arxiv_version","alias_value":"2211.12092v1","created_at":"2026-07-05T05:18:13.001852+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.12092","created_at":"2026-07-05T05:18:13.001852+00:00"},{"alias_kind":"pith_short_12","alias_value":"DM4MZW2ZOGSV","created_at":"2026-07-05T05:18:13.001852+00:00"},{"alias_kind":"pith_short_16","alias_value":"DM4MZW2ZOGSVYQXV","created_at":"2026-07-05T05:18:13.001852+00:00"},{"alias_kind":"pith_short_8","alias_value":"DM4MZW2Z","created_at":"2026-07-05T05:18:13.001852+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.14190","citing_title":"AsyncSwitch: Asynchronous Text-Speech Adaptation for Code-Switched ASR","ref_index":35,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI","json":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI.json","graph_json":"https://pith.science/api/pith-number/DM4MZW2ZOGSVYQXVWAB77I4IFI/graph.json","events_json":"https://pith.science/api/pith-number/DM4MZW2ZOGSVYQXVWAB77I4IFI/events.json","paper":"https://pith.science/paper/DM4MZW2Z"},"agent_actions":{"view_html":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI","download_json":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI.json","view_paper":"https://pith.science/paper/DM4MZW2Z","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2211.12092&json=true","fetch_graph":"https://pith.science/api/pith-number/DM4MZW2ZOGSVYQXVWAB77I4IFI/graph.json","fetch_events":"https://pith.science/api/pith-number/DM4MZW2ZOGSVYQXVWAB77I4IFI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI/action/storage_attestation","attest_author":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI/action/author_attestation","sign_citation":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI/action/citation_signature","submit_replication":"https://pith.science/pith/DM4MZW2ZOGSVYQXVWAB77I4IFI/action/replication_record"}},"created_at":"2026-07-05T05:18:13.001852+00:00","updated_at":"2026-07-05T05:18:13.001852+00:00"}