{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4W7LMABFUQLNRGPCPDSGR4ACET","short_pith_number":"pith:4W7LMABF","schema_version":"1.0","canonical_sha256":"e5beb60025a416d899e278e468f00224cc17c945f8963006ea6a9bb6c513e296","source":{"kind":"arxiv","id":"2404.15737","version":2},"attestation_state":"computed","paper":{"title":"No Train but Gain: Language Arithmetic for training-free Language Adapters enhancement","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexandra Birch, Mateusz Klimaszewski, Piotr Andruszkiewicz","submitted_at":"2024-04-24T08:52:40Z","abstract_excerpt":"Modular deep learning is the state-of-the-art solution for lifting the curse of multilinguality, preventing the impact of negative interference and enabling cross-lingual performance in Multilingual Pre-trained Language Models. However, a trade-off of this approach is the reduction in positive transfer learning from closely related languages. In response, we introduce a novel method called language arithmetic, which enables training-free post-processing to address this limitation. Extending the task arithmetic framework, we apply learning via addition to the language adapters, transitioning th"},"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":"2404.15737","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-24T08:52:40Z","cross_cats_sorted":[],"title_canon_sha256":"82c3fe026add79ba9ca8b1ac9c43b09b4eb83034a6a262e7613f0f93be2e446b","abstract_canon_sha256":"79aae6c17a77a7c8f391564e358ee36bfb366dca7f146503dc26a882da3556d1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:04:35.377136Z","signature_b64":"Rk+81cX7ZfJTWtPikUTp3ml4PxjN8q386MdZM4Pt8Sytw5qktddQUpJo/UANkAdWkXmGSP41AMi7LC1UY41DDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e5beb60025a416d899e278e468f00224cc17c945f8963006ea6a9bb6c513e296","last_reissued_at":"2026-07-05T09:04:35.376645Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:04:35.376645Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"No Train but Gain: Language Arithmetic for training-free Language Adapters enhancement","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexandra Birch, Mateusz Klimaszewski, Piotr Andruszkiewicz","submitted_at":"2024-04-24T08:52:40Z","abstract_excerpt":"Modular deep learning is the state-of-the-art solution for lifting the curse of multilinguality, preventing the impact of negative interference and enabling cross-lingual performance in Multilingual Pre-trained Language Models. However, a trade-off of this approach is the reduction in positive transfer learning from closely related languages. In response, we introduce a novel method called language arithmetic, which enables training-free post-processing to address this limitation. Extending the task arithmetic framework, we apply learning via addition to the language adapters, transitioning th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.15737","kind":"arxiv","version":2},"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/2404.15737/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":"2404.15737","created_at":"2026-07-05T09:04:35.376706+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.15737v2","created_at":"2026-07-05T09:04:35.376706+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.15737","created_at":"2026-07-05T09:04:35.376706+00:00"},{"alias_kind":"pith_short_12","alias_value":"4W7LMABFUQLN","created_at":"2026-07-05T09:04:35.376706+00:00"},{"alias_kind":"pith_short_16","alias_value":"4W7LMABFUQLNRGPC","created_at":"2026-07-05T09:04:35.376706+00:00"},{"alias_kind":"pith_short_8","alias_value":"4W7LMABF","created_at":"2026-07-05T09:04:35.376706+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2408.01119","citing_title":"Task Prompt Vectors: Effective Initialization through Multi-Task Soft-Prompt Transfer","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2408.07666","citing_title":"Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities","ref_index":111,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET","json":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET.json","graph_json":"https://pith.science/api/pith-number/4W7LMABFUQLNRGPCPDSGR4ACET/graph.json","events_json":"https://pith.science/api/pith-number/4W7LMABFUQLNRGPCPDSGR4ACET/events.json","paper":"https://pith.science/paper/4W7LMABF"},"agent_actions":{"view_html":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET","download_json":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET.json","view_paper":"https://pith.science/paper/4W7LMABF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.15737&json=true","fetch_graph":"https://pith.science/api/pith-number/4W7LMABFUQLNRGPCPDSGR4ACET/graph.json","fetch_events":"https://pith.science/api/pith-number/4W7LMABFUQLNRGPCPDSGR4ACET/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET/action/storage_attestation","attest_author":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET/action/author_attestation","sign_citation":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET/action/citation_signature","submit_replication":"https://pith.science/pith/4W7LMABFUQLNRGPCPDSGR4ACET/action/replication_record"}},"created_at":"2026-07-05T09:04:35.376706+00:00","updated_at":"2026-07-05T09:04:35.376706+00:00"}