{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FEGUURDYEV2LZMW6UIXYHFDXIR","short_pith_number":"pith:FEGUURDY","schema_version":"1.0","canonical_sha256":"290d4a44782574bcb2dea22f839477446c8ff11df062a6d76d39ccb6585aa780","source":{"kind":"arxiv","id":"2109.07740","version":1},"attestation_state":"computed","paper":{"title":"Scaling Laws for Neural Machine Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Ankur Bapna, Behrooz Ghorbani, Ciprian Chelba, Colin Cherry, Markus Freitag, Maxim Krikun, Orhan Firat, Xavier Garcia","submitted_at":"2021-09-16T06:15:20Z","abstract_excerpt":"We present an empirical study of scaling properties of encoder-decoder Transformer models used in neural machine translation (NMT). We show that cross-entropy loss as a function of model size follows a certain scaling law. Specifically (i) We propose a formula which describes the scaling behavior of cross-entropy loss as a bivariate function of encoder and decoder size, and show that it gives accurate predictions under a variety of scaling approaches and languages; we show that the total number of parameters alone is not sufficient for such purposes. (ii) We observe different power law exponen"},"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":"2109.07740","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-09-16T06:15:20Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"83481e6271e65f34cd7c0dd3a83e5c8f957afb57c425006e663d22d4a31805ab","abstract_canon_sha256":"9019202902466df6c3c42a10dfd67f4986540a7080082aac3bf33f65e091b194"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:15:00.520924Z","signature_b64":"+hgmZyFCXt87JQx5BT98XoYW6G/Pw5ApCq8yNeSKzo7FR4EUvGv6pFmE+WI6TJgYYHWpnDorZhxMCRQft9kcCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"290d4a44782574bcb2dea22f839477446c8ff11df062a6d76d39ccb6585aa780","last_reissued_at":"2026-07-05T03:15:00.520492Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:15:00.520492Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Scaling Laws for Neural Machine Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Ankur Bapna, Behrooz Ghorbani, Ciprian Chelba, Colin Cherry, Markus Freitag, Maxim Krikun, Orhan Firat, Xavier Garcia","submitted_at":"2021-09-16T06:15:20Z","abstract_excerpt":"We present an empirical study of scaling properties of encoder-decoder Transformer models used in neural machine translation (NMT). We show that cross-entropy loss as a function of model size follows a certain scaling law. Specifically (i) We propose a formula which describes the scaling behavior of cross-entropy loss as a bivariate function of encoder and decoder size, and show that it gives accurate predictions under a variety of scaling approaches and languages; we show that the total number of parameters alone is not sufficient for such purposes. (ii) We observe different power law exponen"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.07740","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/2109.07740/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":"2109.07740","created_at":"2026-07-05T03:15:00.520549+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.07740v1","created_at":"2026-07-05T03:15:00.520549+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.07740","created_at":"2026-07-05T03:15:00.520549+00:00"},{"alias_kind":"pith_short_12","alias_value":"FEGUURDYEV2L","created_at":"2026-07-05T03:15:00.520549+00:00"},{"alias_kind":"pith_short_16","alias_value":"FEGUURDYEV2LZMW6","created_at":"2026-07-05T03:15:00.520549+00:00"},{"alias_kind":"pith_short_8","alias_value":"FEGUURDY","created_at":"2026-07-05T03:15:00.520549+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.27989","citing_title":"Law of Neural Interaction: Depth-Width Shape, Interaction Efficiency, and Generalization","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2405.00592","citing_title":"Scaling and renormalization in high-dimensional regression","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2502.05074","citing_title":"Two-Point Deterministic Equivalence for Stochastic Gradient Dynamics in Linear Models","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2305.16264","citing_title":"Scaling Data-Constrained Language Models","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2405.14782","citing_title":"Lessons from the Trenches on Reproducible Evaluation of Language Models","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2304.01373","citing_title":"Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling","ref_index":124,"is_internal_anchor":false},{"citing_arxiv_id":"2308.08998","citing_title":"Reinforced Self-Training (ReST) for Language Modeling","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR","json":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR.json","graph_json":"https://pith.science/api/pith-number/FEGUURDYEV2LZMW6UIXYHFDXIR/graph.json","events_json":"https://pith.science/api/pith-number/FEGUURDYEV2LZMW6UIXYHFDXIR/events.json","paper":"https://pith.science/paper/FEGUURDY"},"agent_actions":{"view_html":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR","download_json":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR.json","view_paper":"https://pith.science/paper/FEGUURDY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.07740&json=true","fetch_graph":"https://pith.science/api/pith-number/FEGUURDYEV2LZMW6UIXYHFDXIR/graph.json","fetch_events":"https://pith.science/api/pith-number/FEGUURDYEV2LZMW6UIXYHFDXIR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR/action/storage_attestation","attest_author":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR/action/author_attestation","sign_citation":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR/action/citation_signature","submit_replication":"https://pith.science/pith/FEGUURDYEV2LZMW6UIXYHFDXIR/action/replication_record"}},"created_at":"2026-07-05T03:15:00.520549+00:00","updated_at":"2026-07-05T03:15:00.520549+00:00"}