{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4KWDXAUGYITLMEFIR6UA34VAQO","short_pith_number":"pith:4KWDXAUG","schema_version":"1.0","canonical_sha256":"e2ac3b8286c226b610a88fa80df2a0839e0e2205a65b6e35b73721ff80110e42","source":{"kind":"arxiv","id":"2412.07633","version":1},"attestation_state":"computed","paper":{"title":"ChocoLlama: Lessons Learned From Teaching Llamas Dutch","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anthony Rath\\'e, Fran\\c{c}ois Remy, Jens-Joris Decorte, Matthieu Meeus, Pieter Delobelle, Thomas Demeester","submitted_at":"2024-12-10T16:13:58Z","abstract_excerpt":"While Large Language Models (LLMs) have shown remarkable capabilities in natural language understanding and generation, their performance often lags in lower-resource, non-English languages due to biases in the training data. In this work, we explore strategies for adapting the primarily English LLMs (Llama-2 and Llama-3) to Dutch, a language spoken by 30 million people worldwide yet often underrepresented in LLM development. We collect 104GB of Dutch text ($32$B tokens) from various sources to first apply continued pretraining using low-rank adaptation (LoRA), complemented with Dutch posttrai"},"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":"2412.07633","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-12-10T16:13:58Z","cross_cats_sorted":[],"title_canon_sha256":"e3500a86b0d3fe0183af191b1716cae516f306b83f7eb7a0748a85f689cfdaaf","abstract_canon_sha256":"e8dadd4689ad666fc0c1fcd57a56dc46d4fa8cd4c6df7f71613d1e5776b0060c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:47:12.593564Z","signature_b64":"B3OisS795dV/9lMPim+9Wblve7OGjDxDLWeD45W3GCS3UWJwK3IPVAqi9Nw3TtqYr8TWoRq+0ov1cu01WlnkCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e2ac3b8286c226b610a88fa80df2a0839e0e2205a65b6e35b73721ff80110e42","last_reissued_at":"2026-07-05T09:47:12.593056Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:47:12.593056Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ChocoLlama: Lessons Learned From Teaching Llamas Dutch","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anthony Rath\\'e, Fran\\c{c}ois Remy, Jens-Joris Decorte, Matthieu Meeus, Pieter Delobelle, Thomas Demeester","submitted_at":"2024-12-10T16:13:58Z","abstract_excerpt":"While Large Language Models (LLMs) have shown remarkable capabilities in natural language understanding and generation, their performance often lags in lower-resource, non-English languages due to biases in the training data. In this work, we explore strategies for adapting the primarily English LLMs (Llama-2 and Llama-3) to Dutch, a language spoken by 30 million people worldwide yet often underrepresented in LLM development. We collect 104GB of Dutch text ($32$B tokens) from various sources to first apply continued pretraining using low-rank adaptation (LoRA), complemented with Dutch posttrai"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.07633","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/2412.07633/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":"2412.07633","created_at":"2026-07-05T09:47:12.593117+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.07633v1","created_at":"2026-07-05T09:47:12.593117+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.07633","created_at":"2026-07-05T09:47:12.593117+00:00"},{"alias_kind":"pith_short_12","alias_value":"4KWDXAUGYITL","created_at":"2026-07-05T09:47:12.593117+00:00"},{"alias_kind":"pith_short_16","alias_value":"4KWDXAUGYITLMEFI","created_at":"2026-07-05T09:47:12.593117+00:00"},{"alias_kind":"pith_short_8","alias_value":"4KWDXAUG","created_at":"2026-07-05T09:47:12.593117+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07721","citing_title":"Automatic Extraction of Structured Information from Brain MRI Reports Using an Open-Weight Large Language Model","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09645","citing_title":"Generating High Quality Synthetic Data for Dutch Medical Conversations","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO","json":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO.json","graph_json":"https://pith.science/api/pith-number/4KWDXAUGYITLMEFIR6UA34VAQO/graph.json","events_json":"https://pith.science/api/pith-number/4KWDXAUGYITLMEFIR6UA34VAQO/events.json","paper":"https://pith.science/paper/4KWDXAUG"},"agent_actions":{"view_html":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO","download_json":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO.json","view_paper":"https://pith.science/paper/4KWDXAUG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.07633&json=true","fetch_graph":"https://pith.science/api/pith-number/4KWDXAUGYITLMEFIR6UA34VAQO/graph.json","fetch_events":"https://pith.science/api/pith-number/4KWDXAUGYITLMEFIR6UA34VAQO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO/action/storage_attestation","attest_author":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO/action/author_attestation","sign_citation":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO/action/citation_signature","submit_replication":"https://pith.science/pith/4KWDXAUGYITLMEFIR6UA34VAQO/action/replication_record"}},"created_at":"2026-07-05T09:47:12.593117+00:00","updated_at":"2026-07-05T09:47:12.593117+00:00"}