{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:P7NZ2IX325RG5D4QSQMIEEYLGU","short_pith_number":"pith:P7NZ2IX3","schema_version":"1.0","canonical_sha256":"7fdb9d22fbd7626e8f90941882130b353eb68db0d0e1f717fa46c4da4f9d8bbf","source":{"kind":"arxiv","id":"2304.03245","version":3},"attestation_state":"computed","paper":{"title":"Large language models effectively leverage document-level context for literary translation, but critical errors persist","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Marzena Karpinska, Mohit Iyyer","submitted_at":"2023-04-06T17:27:45Z","abstract_excerpt":"Large language models (LLMs) are competitive with the state of the art on a wide range of sentence-level translation datasets. However, their ability to translate paragraphs and documents remains unexplored because evaluation in these settings is costly and difficult. We show through a rigorous human evaluation that asking the Gpt-3.5 (text-davinci-003) LLM to translate an entire literary paragraph (e.g., from a novel) at once results in higher-quality translations than standard sentence-by-sentence translation across 18 linguistically-diverse language pairs (e.g., translating into and out of "},"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":"2304.03245","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-04-06T17:27:45Z","cross_cats_sorted":[],"title_canon_sha256":"877e771ffb24d141a78eebcc6fb903d09b31d552d990db385c65ddf24623bd50","abstract_canon_sha256":"533fa55c05acd7e8b8d3d51785062df7bcacd03db60d17cb7d6f9df8818c643b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:12:47.173341Z","signature_b64":"732knfrS1fUc1TI+JeXr/RqQxtk8bUAQ5sQnSWyASguxXWsWJfgUQnfqyNrCKh4rQh+ARvrAAVU6dBXbBCvCDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7fdb9d22fbd7626e8f90941882130b353eb68db0d0e1f717fa46c4da4f9d8bbf","last_reissued_at":"2026-07-05T06:12:47.172859Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:12:47.172859Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large language models effectively leverage document-level context for literary translation, but critical errors persist","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Marzena Karpinska, Mohit Iyyer","submitted_at":"2023-04-06T17:27:45Z","abstract_excerpt":"Large language models (LLMs) are competitive with the state of the art on a wide range of sentence-level translation datasets. However, their ability to translate paragraphs and documents remains unexplored because evaluation in these settings is costly and difficult. We show through a rigorous human evaluation that asking the Gpt-3.5 (text-davinci-003) LLM to translate an entire literary paragraph (e.g., from a novel) at once results in higher-quality translations than standard sentence-by-sentence translation across 18 linguistically-diverse language pairs (e.g., translating into and out of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.03245","kind":"arxiv","version":3},"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/2304.03245/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":"2304.03245","created_at":"2026-07-05T06:12:47.172914+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.03245v3","created_at":"2026-07-05T06:12:47.172914+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.03245","created_at":"2026-07-05T06:12:47.172914+00:00"},{"alias_kind":"pith_short_12","alias_value":"P7NZ2IX325RG","created_at":"2026-07-05T06:12:47.172914+00:00"},{"alias_kind":"pith_short_16","alias_value":"P7NZ2IX325RG5D4Q","created_at":"2026-07-05T06:12:47.172914+00:00"},{"alias_kind":"pith_short_8","alias_value":"P7NZ2IX3","created_at":"2026-07-05T06:12:47.172914+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.09564","citing_title":"ACE-Bench: A Lightweight Benchmark for Evaluating Azure SDK Usage Correctness","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2412.05579","citing_title":"LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods","ref_index":108,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU","json":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU.json","graph_json":"https://pith.science/api/pith-number/P7NZ2IX325RG5D4QSQMIEEYLGU/graph.json","events_json":"https://pith.science/api/pith-number/P7NZ2IX325RG5D4QSQMIEEYLGU/events.json","paper":"https://pith.science/paper/P7NZ2IX3"},"agent_actions":{"view_html":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU","download_json":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU.json","view_paper":"https://pith.science/paper/P7NZ2IX3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.03245&json=true","fetch_graph":"https://pith.science/api/pith-number/P7NZ2IX325RG5D4QSQMIEEYLGU/graph.json","fetch_events":"https://pith.science/api/pith-number/P7NZ2IX325RG5D4QSQMIEEYLGU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU/action/storage_attestation","attest_author":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU/action/author_attestation","sign_citation":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU/action/citation_signature","submit_replication":"https://pith.science/pith/P7NZ2IX325RG5D4QSQMIEEYLGU/action/replication_record"}},"created_at":"2026-07-05T06:12:47.172914+00:00","updated_at":"2026-07-05T06:12:47.172914+00:00"}