{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QUMZCWWHJ7APQC67ZFZB7DPFS7","short_pith_number":"pith:QUMZCWWH","schema_version":"1.0","canonical_sha256":"8519915ac74fc0f80bdfc9721f8de597d012c594242029c309ffd111f92628d7","source":{"kind":"arxiv","id":"2405.20089","version":2},"attestation_state":"computed","paper":{"title":"The Fine-Tuning Paradox: Boosting Translation Quality Without Sacrificing LLM Abilities","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bill Byrne, Christof Monz, David Stap, Eva Hasler, Ke Tran","submitted_at":"2024-05-30T14:25:56Z","abstract_excerpt":"Fine-tuning large language models (LLMs) for machine translation has shown improvements in overall translation quality. However, it is unclear what is the impact of fine-tuning on desirable LLM behaviors that are not present in neural machine translation models, such as steerability, inherent document-level translation abilities, and the ability to produce less literal translations. We perform an extensive translation evaluation on the LLaMA and Falcon family of models with model size ranging from 7 billion up to 65 billion parameters. Our results show that while fine-tuning improves the gener"},"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":"2405.20089","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-30T14:25:56Z","cross_cats_sorted":[],"title_canon_sha256":"542b49fb3c5ef59d5fcd10ce5d788470ac9ebddc3208ea3a99d662736481aab9","abstract_canon_sha256":"910140b659006d1538424afc769eef77b9f8e1014557372f4d52f240c798f560"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:52:33.524283Z","signature_b64":"3OKEsJpZQgRsnxGf86EZ7fA6lG/eV6cdq3FhL7hanhGinPdECkryhQ2KZFjOGsokIgtzDK6ck6+3/rBtQVNbDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8519915ac74fc0f80bdfc9721f8de597d012c594242029c309ffd111f92628d7","last_reissued_at":"2026-07-05T08:52:33.523860Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:52:33.523860Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Fine-Tuning Paradox: Boosting Translation Quality Without Sacrificing LLM Abilities","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bill Byrne, Christof Monz, David Stap, Eva Hasler, Ke Tran","submitted_at":"2024-05-30T14:25:56Z","abstract_excerpt":"Fine-tuning large language models (LLMs) for machine translation has shown improvements in overall translation quality. However, it is unclear what is the impact of fine-tuning on desirable LLM behaviors that are not present in neural machine translation models, such as steerability, inherent document-level translation abilities, and the ability to produce less literal translations. We perform an extensive translation evaluation on the LLaMA and Falcon family of models with model size ranging from 7 billion up to 65 billion parameters. Our results show that while fine-tuning improves the gener"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.20089","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/2405.20089/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":"2405.20089","created_at":"2026-07-05T08:52:33.523916+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.20089v2","created_at":"2026-07-05T08:52:33.523916+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.20089","created_at":"2026-07-05T08:52:33.523916+00:00"},{"alias_kind":"pith_short_12","alias_value":"QUMZCWWHJ7AP","created_at":"2026-07-05T08:52:33.523916+00:00"},{"alias_kind":"pith_short_16","alias_value":"QUMZCWWHJ7APQC67","created_at":"2026-07-05T08:52:33.523916+00:00"},{"alias_kind":"pith_short_8","alias_value":"QUMZCWWH","created_at":"2026-07-05T08:52:33.523916+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.02995","citing_title":"It's Not a Walk in the Park! Challenges of Idiom Translation in Speech-to-text Systems","ref_index":30,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7","json":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7.json","graph_json":"https://pith.science/api/pith-number/QUMZCWWHJ7APQC67ZFZB7DPFS7/graph.json","events_json":"https://pith.science/api/pith-number/QUMZCWWHJ7APQC67ZFZB7DPFS7/events.json","paper":"https://pith.science/paper/QUMZCWWH"},"agent_actions":{"view_html":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7","download_json":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7.json","view_paper":"https://pith.science/paper/QUMZCWWH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.20089&json=true","fetch_graph":"https://pith.science/api/pith-number/QUMZCWWHJ7APQC67ZFZB7DPFS7/graph.json","fetch_events":"https://pith.science/api/pith-number/QUMZCWWHJ7APQC67ZFZB7DPFS7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7/action/storage_attestation","attest_author":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7/action/author_attestation","sign_citation":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7/action/citation_signature","submit_replication":"https://pith.science/pith/QUMZCWWHJ7APQC67ZFZB7DPFS7/action/replication_record"}},"created_at":"2026-07-05T08:52:33.523916+00:00","updated_at":"2026-07-05T08:52:33.523916+00:00"}