{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:BNIWH4BSQ5YQKRYB4IFYEIAUAV","short_pith_number":"pith:BNIWH4BS","canonical_record":{"source":{"id":"2305.14481","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-23T19:21:53Z","cross_cats_sorted":[],"title_canon_sha256":"1f0379fb453c1aad6611b76de4aae9a3076e9b06f8c654ecf85add12d327b0f0","abstract_canon_sha256":"cb462006c445d44e67438aec56f39d0575587b9e7f1307a78110bcfc7c64a207"},"schema_version":"1.0"},"canonical_sha256":"0b5163f0328771054701e20b822014057f247849b12140c58c298747a655fc50","source":{"kind":"arxiv","id":"2305.14481","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.14481","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"arxiv_version","alias_value":"2305.14481v2","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14481","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"pith_short_12","alias_value":"BNIWH4BSQ5YQ","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"pith_short_16","alias_value":"BNIWH4BSQ5YQKRYB","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"pith_short_8","alias_value":"BNIWH4BS","created_at":"2026-07-05T08:04:12Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:BNIWH4BSQ5YQKRYB4IFYEIAUAV","target":"record","payload":{"canonical_record":{"source":{"id":"2305.14481","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-23T19:21:53Z","cross_cats_sorted":[],"title_canon_sha256":"1f0379fb453c1aad6611b76de4aae9a3076e9b06f8c654ecf85add12d327b0f0","abstract_canon_sha256":"cb462006c445d44e67438aec56f39d0575587b9e7f1307a78110bcfc7c64a207"},"schema_version":"1.0"},"canonical_sha256":"0b5163f0328771054701e20b822014057f247849b12140c58c298747a655fc50","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:04:12.608936Z","signature_b64":"87HFDsqQvM/In1IgwC7hBuvrb9i7xgBq0k+C59mkYy+wcM9luIMcRMiAttx1AhtjdQ0sT/R6XkUwzEnTsey1Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b5163f0328771054701e20b822014057f247849b12140c58c298747a655fc50","last_reissued_at":"2026-07-05T08:04:12.608482Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:04:12.608482Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2305.14481","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:04:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"hJ0HFxJorgpBG7WVsGdzkFxst5vtPJxHUpp44udcDDEtT7TNrBH0O/pUYW7vDxsqOn1B442Esh9XRYeFHEBJCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T23:06:13.561203Z"},"content_sha256":"77491f90381ce02a40f25435fcaef8cf6e5bb149c3a666c132b78e0af7cb0564","schema_version":"1.0","event_id":"sha256:77491f90381ce02a40f25435fcaef8cf6e5bb149c3a666c132b78e0af7cb0564"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:BNIWH4BSQ5YQKRYB4IFYEIAUAV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"FOCUS: Effective Embedding Initialization for Monolingual Specialization of Multilingual Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Gerard de Melo, Konstantin Dobler","submitted_at":"2023-05-23T19:21:53Z","abstract_excerpt":"Using model weights pretrained on a high-resource language as a warm start can reduce the need for data and compute to obtain high-quality language models for other, especially low-resource, languages. However, if we want to use a new tokenizer specialized for the target language, we cannot transfer the source model's embedding matrix. In this paper, we propose FOCUS - Fast Overlapping Token Combinations Using Sparsemax, a novel embedding initialization method that initializes the embedding matrix effectively for a new tokenizer based on information in the source model's embedding matrix. FOCU"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14481","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/2305.14481/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:04:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"36+PRQhPV83JFo972Oojdv2F5VS+zcA0bbdrDodvwUoMB8DJzFxibjh1fkE4ouS+1PinhAZZqqNYbKKm8WrsBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T23:06:13.561735Z"},"content_sha256":"10c8ac0636cc000debffeec31eedb63b3a631bc013572f5f953839b0157b9f3b","schema_version":"1.0","event_id":"sha256:10c8ac0636cc000debffeec31eedb63b3a631bc013572f5f953839b0157b9f3b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BNIWH4BSQ5YQKRYB4IFYEIAUAV/bundle.json","state_url":"https://pith.science/pith/BNIWH4BSQ5YQKRYB4IFYEIAUAV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BNIWH4BSQ5YQKRYB4IFYEIAUAV/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-04T23:06:13Z","links":{"resolver":"https://pith.science/pith/BNIWH4BSQ5YQKRYB4IFYEIAUAV","bundle":"https://pith.science/pith/BNIWH4BSQ5YQKRYB4IFYEIAUAV/bundle.json","state":"https://pith.science/pith/BNIWH4BSQ5YQKRYB4IFYEIAUAV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BNIWH4BSQ5YQKRYB4IFYEIAUAV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:BNIWH4BSQ5YQKRYB4IFYEIAUAV","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"cb462006c445d44e67438aec56f39d0575587b9e7f1307a78110bcfc7c64a207","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-23T19:21:53Z","title_canon_sha256":"1f0379fb453c1aad6611b76de4aae9a3076e9b06f8c654ecf85add12d327b0f0"},"schema_version":"1.0","source":{"id":"2305.14481","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.14481","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"arxiv_version","alias_value":"2305.14481v2","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14481","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"pith_short_12","alias_value":"BNIWH4BSQ5YQ","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"pith_short_16","alias_value":"BNIWH4BSQ5YQKRYB","created_at":"2026-07-05T08:04:12Z"},{"alias_kind":"pith_short_8","alias_value":"BNIWH4BS","created_at":"2026-07-05T08:04:12Z"}],"graph_snapshots":[{"event_id":"sha256:10c8ac0636cc000debffeec31eedb63b3a631bc013572f5f953839b0157b9f3b","target":"graph","created_at":"2026-07-05T08:04:12Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2305.14481/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Using model weights pretrained on a high-resource language as a warm start can reduce the need for data and compute to obtain high-quality language models for other, especially low-resource, languages. However, if we want to use a new tokenizer specialized for the target language, we cannot transfer the source model's embedding matrix. In this paper, we propose FOCUS - Fast Overlapping Token Combinations Using Sparsemax, a novel embedding initialization method that initializes the embedding matrix effectively for a new tokenizer based on information in the source model's embedding matrix. FOCU","authors_text":"Gerard de Melo, Konstantin Dobler","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-23T19:21:53Z","title":"FOCUS: Effective Embedding Initialization for Monolingual Specialization of Multilingual Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14481","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:77491f90381ce02a40f25435fcaef8cf6e5bb149c3a666c132b78e0af7cb0564","target":"record","created_at":"2026-07-05T08:04:12Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"cb462006c445d44e67438aec56f39d0575587b9e7f1307a78110bcfc7c64a207","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-23T19:21:53Z","title_canon_sha256":"1f0379fb453c1aad6611b76de4aae9a3076e9b06f8c654ecf85add12d327b0f0"},"schema_version":"1.0","source":{"id":"2305.14481","kind":"arxiv","version":2}},"canonical_sha256":"0b5163f0328771054701e20b822014057f247849b12140c58c298747a655fc50","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0b5163f0328771054701e20b822014057f247849b12140c58c298747a655fc50","first_computed_at":"2026-07-05T08:04:12.608482Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:04:12.608482Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"87HFDsqQvM/In1IgwC7hBuvrb9i7xgBq0k+C59mkYy+wcM9luIMcRMiAttx1AhtjdQ0sT/R6XkUwzEnTsey1Ag==","signature_status":"signed_v1","signed_at":"2026-07-05T08:04:12.608936Z","signed_message":"canonical_sha256_bytes"},"source_id":"2305.14481","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:77491f90381ce02a40f25435fcaef8cf6e5bb149c3a666c132b78e0af7cb0564","sha256:10c8ac0636cc000debffeec31eedb63b3a631bc013572f5f953839b0157b9f3b"],"state_sha256":"d0f3880a5b45247ef74df4225c75565ff8455ff74483efa01df7a4517cfb5628"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TOLNzXw/JMUxVGB/wrsSPsJlI99WzXjIJqydsQIR8KArA6ouWACIl6xcAeOf26RssT9zf4MMc7TsidSwMp1/BA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T23:06:13.565590Z","bundle_sha256":"f50fcc45c069dfd694683ebec5a9d68ca67c59bdc5248b2dc1d02b1bdee1bcc7"}}