{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:JPHODUCSKUMHNSHBLTT46KDPCL","short_pith_number":"pith:JPHODUCS","schema_version":"1.0","canonical_sha256":"4bcee1d052551876c8e15ce7cf286f12dfb49d595d2a992d396e890fb488140a","source":{"kind":"arxiv","id":"2401.01854","version":4},"attestation_state":"computed","paper":{"title":"Multilingual Instruction Tuning With Just a Pinch of Multilinguality","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Idan Szpektor, Jonathan Herzig, Matan Eyal, Reut Tsarfaty, Roee Aharoni, Uri Shaham","submitted_at":"2024-01-03T17:48:10Z","abstract_excerpt":"As instruction-tuned large language models (LLMs) gain global adoption, their ability to follow instructions in multiple languages becomes increasingly crucial. In this work, we investigate how multilinguality during instruction tuning of a multilingual LLM affects instruction-following across languages from the pre-training corpus. We first show that many languages transfer some instruction-following capabilities to other languages from even monolingual tuning. Furthermore, we find that only 40 multilingual examples integrated in an English tuning set substantially improve multilingual instru"},"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":"2401.01854","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-01-03T17:48:10Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"5c1c3b9ceee012a193184ad4f1f7dd9c91a5b4a9d5619f586e72eaecec1a0daa","abstract_canon_sha256":"64254d6dce81df8a13284e8abf9bceb97062831ba5c622733eb6d1eb97456e24"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:21:11.027211Z","signature_b64":"4ZRyE9zH1aY9IPtPUhgNsj0V6MQeqNyy3xl4wZl1bieGt8LARAC5oc3tDnuz0sXkndYNaGQBrEMOVNa0YtcDBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4bcee1d052551876c8e15ce7cf286f12dfb49d595d2a992d396e890fb488140a","last_reissued_at":"2026-07-05T08:21:11.026766Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:21:11.026766Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multilingual Instruction Tuning With Just a Pinch of Multilinguality","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Idan Szpektor, Jonathan Herzig, Matan Eyal, Reut Tsarfaty, Roee Aharoni, Uri Shaham","submitted_at":"2024-01-03T17:48:10Z","abstract_excerpt":"As instruction-tuned large language models (LLMs) gain global adoption, their ability to follow instructions in multiple languages becomes increasingly crucial. In this work, we investigate how multilinguality during instruction tuning of a multilingual LLM affects instruction-following across languages from the pre-training corpus. We first show that many languages transfer some instruction-following capabilities to other languages from even monolingual tuning. Furthermore, we find that only 40 multilingual examples integrated in an English tuning set substantially improve multilingual instru"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.01854","kind":"arxiv","version":4},"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/2401.01854/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":"2401.01854","created_at":"2026-07-05T08:21:11.026823+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.01854v4","created_at":"2026-07-05T08:21:11.026823+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.01854","created_at":"2026-07-05T08:21:11.026823+00:00"},{"alias_kind":"pith_short_12","alias_value":"JPHODUCSKUMH","created_at":"2026-07-05T08:21:11.026823+00:00"},{"alias_kind":"pith_short_16","alias_value":"JPHODUCSKUMHNSHB","created_at":"2026-07-05T08:21:11.026823+00:00"},{"alias_kind":"pith_short_8","alias_value":"JPHODUCS","created_at":"2026-07-05T08:21:11.026823+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.24114","citing_title":"IRIS: Interleaved Reinforcement with Incremental Staged Curriculum for Cross-Lingual Mathematical Reasoning","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13286","citing_title":"English is Not All You Need: Systematically Exploring the Role of Multilinguality in LLM Post-Training","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL","json":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL.json","graph_json":"https://pith.science/api/pith-number/JPHODUCSKUMHNSHBLTT46KDPCL/graph.json","events_json":"https://pith.science/api/pith-number/JPHODUCSKUMHNSHBLTT46KDPCL/events.json","paper":"https://pith.science/paper/JPHODUCS"},"agent_actions":{"view_html":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL","download_json":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL.json","view_paper":"https://pith.science/paper/JPHODUCS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.01854&json=true","fetch_graph":"https://pith.science/api/pith-number/JPHODUCSKUMHNSHBLTT46KDPCL/graph.json","fetch_events":"https://pith.science/api/pith-number/JPHODUCSKUMHNSHBLTT46KDPCL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL/action/storage_attestation","attest_author":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL/action/author_attestation","sign_citation":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL/action/citation_signature","submit_replication":"https://pith.science/pith/JPHODUCSKUMHNSHBLTT46KDPCL/action/replication_record"}},"created_at":"2026-07-05T08:21:11.026823+00:00","updated_at":"2026-07-05T08:21:11.026823+00:00"}