{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:K3NXAOXTW6BOAO2BWNVHIE4NED","short_pith_number":"pith:K3NXAOXT","schema_version":"1.0","canonical_sha256":"56db703af3b782e03b41b36a74138d20de932513997387d2becd5315abe6fcd4","source":{"kind":"arxiv","id":"2503.10494","version":1},"attestation_state":"computed","paper":{"title":"Source-primed Multi-turn Conversation Helps Large Language Models Translate Documents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hanxu Hu, Jannis Vamvas, Rico Sennrich","submitted_at":"2025-03-13T15:57:50Z","abstract_excerpt":"LLMs have paved the way for truly simple document-level machine translation, but challenges such as omission errors remain. In this paper, we study a simple method for handling document-level machine translation, by leveraging previous contexts in a multi-turn conversational manner. Specifically, by decomposing documents into segments and iteratively translating them while maintaining previous turns, this method ensures coherent translations without additional training, and can fully re-use the KV cache of previous turns thus minimizing computational overhead. We further propose a `source-prim"},"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":"2503.10494","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-13T15:57:50Z","cross_cats_sorted":[],"title_canon_sha256":"b3381691a4804435b001fd9dc8fe03681e80fc928811fa340ec945b2415d808e","abstract_canon_sha256":"af5cdd2ed48730335420c272d5195ae5b07b0fb481ba36b1f90cf1de0db7652a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:30:46.526133Z","signature_b64":"TEOSDBf2EFMxwf3PnueTGjePrwaluhY+1YxFdDwX2pVOGCUoILktD6mOQVjD1/n8WrdXfI8NBHdKywvqjsZ7Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"56db703af3b782e03b41b36a74138d20de932513997387d2becd5315abe6fcd4","last_reissued_at":"2026-07-05T10:30:46.525656Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:30:46.525656Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Source-primed Multi-turn Conversation Helps Large Language Models Translate Documents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hanxu Hu, Jannis Vamvas, Rico Sennrich","submitted_at":"2025-03-13T15:57:50Z","abstract_excerpt":"LLMs have paved the way for truly simple document-level machine translation, but challenges such as omission errors remain. In this paper, we study a simple method for handling document-level machine translation, by leveraging previous contexts in a multi-turn conversational manner. Specifically, by decomposing documents into segments and iteratively translating them while maintaining previous turns, this method ensures coherent translations without additional training, and can fully re-use the KV cache of previous turns thus minimizing computational overhead. We further propose a `source-prim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.10494","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/2503.10494/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":"2503.10494","created_at":"2026-07-05T10:30:46.525716+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.10494v1","created_at":"2026-07-05T10:30:46.525716+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.10494","created_at":"2026-07-05T10:30:46.525716+00:00"},{"alias_kind":"pith_short_12","alias_value":"K3NXAOXTW6BO","created_at":"2026-07-05T10:30:46.525716+00:00"},{"alias_kind":"pith_short_16","alias_value":"K3NXAOXTW6BOAO2B","created_at":"2026-07-05T10:30:46.525716+00:00"},{"alias_kind":"pith_short_8","alias_value":"K3NXAOXT","created_at":"2026-07-05T10:30:46.525716+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.07583","citing_title":"Beyond the Sentence: A Survey on Context-Aware Machine Translation with Large Language Models","ref_index":37,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED","json":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED.json","graph_json":"https://pith.science/api/pith-number/K3NXAOXTW6BOAO2BWNVHIE4NED/graph.json","events_json":"https://pith.science/api/pith-number/K3NXAOXTW6BOAO2BWNVHIE4NED/events.json","paper":"https://pith.science/paper/K3NXAOXT"},"agent_actions":{"view_html":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED","download_json":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED.json","view_paper":"https://pith.science/paper/K3NXAOXT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.10494&json=true","fetch_graph":"https://pith.science/api/pith-number/K3NXAOXTW6BOAO2BWNVHIE4NED/graph.json","fetch_events":"https://pith.science/api/pith-number/K3NXAOXTW6BOAO2BWNVHIE4NED/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED/action/timestamp_anchor","attest_storage":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED/action/storage_attestation","attest_author":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED/action/author_attestation","sign_citation":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED/action/citation_signature","submit_replication":"https://pith.science/pith/K3NXAOXTW6BOAO2BWNVHIE4NED/action/replication_record"}},"created_at":"2026-07-05T10:30:46.525716+00:00","updated_at":"2026-07-05T10:30:46.525716+00:00"}