{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WQ3MBGKQNBC3HINPMHCPFO7HXP","short_pith_number":"pith:WQ3MBGKQ","schema_version":"1.0","canonical_sha256":"b436c099506845b3a1af61c4f2bbe7bbf7dc94eaf82729b430158303d1b2072b","source":{"kind":"arxiv","id":"2604.17633","version":2},"attestation_state":"computed","paper":{"title":"Copy First, Translate Later: Interpreting Translation Dynamics in Multilingual Pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"In multilingual pretraining, models first copy tokens before developing general translation abilities.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Barbara Plank, Felicia K\\\"orner, Florian Eichin, Gitta Kutyniok, Maria Matveev, Michael A. Hedderich","submitted_at":"2026-04-19T22:03:29Z","abstract_excerpt":"Large language models exhibit impressive cross-lingual capabilities. However, prior work analyzes this phenomenon through isolated factors and at sparse points during training, limiting our understanding of how cross-lingual generalization emerges--particularly in the early phases of learning. To study the early trajectory of linguistic and translation capabilities, we pretrain a multilingual 1.7B model on nine diverse languages, capturing checkpoints at a much finer granularity. We use word-level translation as a testbed, introducing a novel dataset to trace how translation develops over trai"},"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":"2604.17633","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-04-19T22:03:29Z","cross_cats_sorted":[],"title_canon_sha256":"bf9a95b979750584e07a21384d86edfc4ed095fff75093558bfb7eecabe4580f","abstract_canon_sha256":"26ecc0725eb1d5d23b9592e66c2c085489973cc5e00cd72c346a18720ea05657"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-29T01:14:31.302372Z","signature_b64":"kdGRIhuIp6kEVKV+uUHFo/622MlJu+VfrmEask4DHUIZWC8Bsns2+wUN68malsb+TKk8bofYg7MYDY+kaB2yAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b436c099506845b3a1af61c4f2bbe7bbf7dc94eaf82729b430158303d1b2072b","last_reissued_at":"2026-06-29T01:14:31.301956Z","signature_status":"signed_v1","first_computed_at":"2026-06-29T01:14:31.301956Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Copy First, Translate Later: Interpreting Translation Dynamics in Multilingual Pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"In multilingual pretraining, models first copy tokens before developing general translation abilities.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Barbara Plank, Felicia K\\\"orner, Florian Eichin, Gitta Kutyniok, Maria Matveev, Michael A. Hedderich","submitted_at":"2026-04-19T22:03:29Z","abstract_excerpt":"Large language models exhibit impressive cross-lingual capabilities. However, prior work analyzes this phenomenon through isolated factors and at sparse points during training, limiting our understanding of how cross-lingual generalization emerges--particularly in the early phases of learning. To study the early trajectory of linguistic and translation capabilities, we pretrain a multilingual 1.7B model on nine diverse languages, capturing checkpoints at a much finer granularity. We use word-level translation as a testbed, introducing a novel dataset to trace how translation develops over trai"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We find that the model quickly acquires basic linguistic capabilities in parallel with token-level copying, while translation develops in two distinct phases: an initial phase dominated by copying and surface-level similarities, and a second phase in which more generalizing translation mechanisms are developed while copying is refined.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The behavioral analyses, model-component inspections, and parameter-based ablations on the chosen checkpoints and novel dataset accurately isolate translation dynamics from confounding factors such as data overlap or model architecture specifics.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Multilingual pretraining develops translation in two phases: early copying driven by surface similarities, followed by generalizing mechanisms while copying is refined.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"In multilingual pretraining, models first copy tokens before developing general translation abilities.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d124692a45724e99173aa811e49d23a336ca79d373a5e9e95a5db13da97201ff"},"source":{"id":"2604.17633","kind":"arxiv","version":2},"verdict":{"id":"d6b52f51-0fc2-47f2-a353-e7cb68dd237d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T05:33:30.283524Z","strongest_claim":"We find that the model quickly acquires basic linguistic capabilities in parallel with token-level copying, while translation develops in two distinct phases: an initial phase dominated by copying and surface-level similarities, and a second phase in which more generalizing translation mechanisms are developed while copying is refined.","one_line_summary":"Multilingual pretraining develops translation in two phases: early copying driven by surface similarities, followed by generalizing mechanisms while copying is refined.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The behavioral analyses, model-component inspections, and parameter-based ablations on the chosen checkpoints and novel dataset accurately isolate translation dynamics from confounding factors such as data overlap or model architecture specifics.","pith_extraction_headline":"In multilingual pretraining, models first copy tokens before developing general translation abilities."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.17633/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":"2604.17633","created_at":"2026-06-29T01:14:31.302015+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.17633v2","created_at":"2026-06-29T01:14:31.302015+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.17633","created_at":"2026-06-29T01:14:31.302015+00:00"},{"alias_kind":"pith_short_12","alias_value":"WQ3MBGKQNBC3","created_at":"2026-06-29T01:14:31.302015+00:00"},{"alias_kind":"pith_short_16","alias_value":"WQ3MBGKQNBC3HINP","created_at":"2026-06-29T01:14:31.302015+00:00"},{"alias_kind":"pith_short_8","alias_value":"WQ3MBGKQ","created_at":"2026-06-29T01:14:31.302015+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/WQ3MBGKQNBC3HINPMHCPFO7HXP","json":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP.json","graph_json":"https://pith.science/api/pith-number/WQ3MBGKQNBC3HINPMHCPFO7HXP/graph.json","events_json":"https://pith.science/api/pith-number/WQ3MBGKQNBC3HINPMHCPFO7HXP/events.json","paper":"https://pith.science/paper/WQ3MBGKQ"},"agent_actions":{"view_html":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP","download_json":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP.json","view_paper":"https://pith.science/paper/WQ3MBGKQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.17633&json=true","fetch_graph":"https://pith.science/api/pith-number/WQ3MBGKQNBC3HINPMHCPFO7HXP/graph.json","fetch_events":"https://pith.science/api/pith-number/WQ3MBGKQNBC3HINPMHCPFO7HXP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP/action/storage_attestation","attest_author":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP/action/author_attestation","sign_citation":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP/action/citation_signature","submit_replication":"https://pith.science/pith/WQ3MBGKQNBC3HINPMHCPFO7HXP/action/replication_record"}},"created_at":"2026-06-29T01:14:31.302015+00:00","updated_at":"2026-06-29T01:14:31.302015+00:00"}