{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:WY6BBASFJLOKSMDE2WBTRH7Y3S","short_pith_number":"pith:WY6BBASF","schema_version":"1.0","canonical_sha256":"b63c1082454adca93064d583389ff8dcb9d67ff571eb95246d223153e1329ba7","source":{"kind":"arxiv","id":"2106.02834","version":1},"attestation_state":"computed","paper":{"title":"MergeDistill: Merging Pre-trained Language Models using Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Melvin Johnson, Partha Talukdar, Simran Khanuja","submitted_at":"2021-06-05T08:22:05Z","abstract_excerpt":"Pre-trained multilingual language models (LMs) have achieved state-of-the-art results in cross-lingual transfer, but they often lead to an inequitable representation of languages due to limited capacity, skewed pre-training data, and sub-optimal vocabularies. This has prompted the creation of an ever-growing pre-trained model universe, where each model is trained on large amounts of language or domain specific data with a carefully curated, linguistically informed vocabulary. However, doing so brings us back full circle and prevents one from leveraging the benefits of multilinguality. To addre"},"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":"2106.02834","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-06-05T08:22:05Z","cross_cats_sorted":[],"title_canon_sha256":"a36fcc0e37edacd15b869ecac84527b9b4210406f946083a4ec47cf61141028b","abstract_canon_sha256":"efd8b93df8290e8cc472bf1c7d611a7e63cc26f5069154c74bd6e8fba3a797d9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:46:30.703939Z","signature_b64":"ZD7krrLW6ZBYPad+Vb1HXx9/oEfBGmGklOJkV+PJLDNGVPV0+Y0uRlPgWfr3xr8prACapmaYycRoYwfXttrBBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b63c1082454adca93064d583389ff8dcb9d67ff571eb95246d223153e1329ba7","last_reissued_at":"2026-07-05T02:46:30.703539Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:46:30.703539Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MergeDistill: Merging Pre-trained Language Models using Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Melvin Johnson, Partha Talukdar, Simran Khanuja","submitted_at":"2021-06-05T08:22:05Z","abstract_excerpt":"Pre-trained multilingual language models (LMs) have achieved state-of-the-art results in cross-lingual transfer, but they often lead to an inequitable representation of languages due to limited capacity, skewed pre-training data, and sub-optimal vocabularies. This has prompted the creation of an ever-growing pre-trained model universe, where each model is trained on large amounts of language or domain specific data with a carefully curated, linguistically informed vocabulary. However, doing so brings us back full circle and prevents one from leveraging the benefits of multilinguality. To addre"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.02834","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/2106.02834/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":"2106.02834","created_at":"2026-07-05T02:46:30.703597+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.02834v1","created_at":"2026-07-05T02:46:30.703597+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.02834","created_at":"2026-07-05T02:46:30.703597+00:00"},{"alias_kind":"pith_short_12","alias_value":"WY6BBASFJLOK","created_at":"2026-07-05T02:46:30.703597+00:00"},{"alias_kind":"pith_short_16","alias_value":"WY6BBASFJLOKSMDE","created_at":"2026-07-05T02:46:30.703597+00:00"},{"alias_kind":"pith_short_8","alias_value":"WY6BBASF","created_at":"2026-07-05T02:46:30.703597+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.04567","citing_title":"StatsMerging: Statistics-Guided Model Merging via Task-Specific Teacher Distillation","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S","json":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S.json","graph_json":"https://pith.science/api/pith-number/WY6BBASFJLOKSMDE2WBTRH7Y3S/graph.json","events_json":"https://pith.science/api/pith-number/WY6BBASFJLOKSMDE2WBTRH7Y3S/events.json","paper":"https://pith.science/paper/WY6BBASF"},"agent_actions":{"view_html":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S","download_json":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S.json","view_paper":"https://pith.science/paper/WY6BBASF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.02834&json=true","fetch_graph":"https://pith.science/api/pith-number/WY6BBASFJLOKSMDE2WBTRH7Y3S/graph.json","fetch_events":"https://pith.science/api/pith-number/WY6BBASFJLOKSMDE2WBTRH7Y3S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S/action/storage_attestation","attest_author":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S/action/author_attestation","sign_citation":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S/action/citation_signature","submit_replication":"https://pith.science/pith/WY6BBASFJLOKSMDE2WBTRH7Y3S/action/replication_record"}},"created_at":"2026-07-05T02:46:30.703597+00:00","updated_at":"2026-07-05T02:46:30.703597+00:00"}