{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:MM67ROAGQZL4RVLWRCE5IHMZL3","short_pith_number":"pith:MM67ROAG","schema_version":"1.0","canonical_sha256":"633df8b8068657c8d5768889d41d995ee730cc292cac1b45cadb005c776cfc68","source":{"kind":"arxiv","id":"2109.07684","version":1},"attestation_state":"computed","paper":{"title":"Language Models are Few-shot Multilingual Learners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Andrea Madotto, Genta Indra Winata, Jason Yosinski, Pascale Fung, Rosanne Liu, Zhaojiang Lin","submitted_at":"2021-09-16T03:08:22Z","abstract_excerpt":"General-purpose language models have demonstrated impressive capabilities, performing on par with state-of-the-art approaches on a range of downstream natural language processing (NLP) tasks and benchmarks when inferring instructions from very few examples. Here, we evaluate the multilingual skills of the GPT and T5 models in conducting multi-class classification on non-English languages without any parameter updates. We show that, given a few English examples as context, pre-trained language models can predict not only English test samples but also non-English ones. Finally, we find the in-co"},"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":"2109.07684","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-09-16T03:08:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1ef7a152403f3fbd827a3186831be6dfc68747db1c83be3383b7afaf1b2bd859","abstract_canon_sha256":"88b1c95d36fdbb8c8ba40a092c18701f246e8868d40f1d4057797b9ef10ccd9e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:14:59.926816Z","signature_b64":"BIcqSKdXhDOBC0NFx9CZsoOwpTVGnyl/UOqi0LnFtE2eYdSUivCZJoMhRp1ah2kYuPUKNB93m9EglH2rOEmPBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"633df8b8068657c8d5768889d41d995ee730cc292cac1b45cadb005c776cfc68","last_reissued_at":"2026-07-05T03:14:59.926367Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:14:59.926367Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language Models are Few-shot Multilingual Learners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Andrea Madotto, Genta Indra Winata, Jason Yosinski, Pascale Fung, Rosanne Liu, Zhaojiang Lin","submitted_at":"2021-09-16T03:08:22Z","abstract_excerpt":"General-purpose language models have demonstrated impressive capabilities, performing on par with state-of-the-art approaches on a range of downstream natural language processing (NLP) tasks and benchmarks when inferring instructions from very few examples. Here, we evaluate the multilingual skills of the GPT and T5 models in conducting multi-class classification on non-English languages without any parameter updates. We show that, given a few English examples as context, pre-trained language models can predict not only English test samples but also non-English ones. Finally, we find the in-co"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.07684","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/2109.07684/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":"2109.07684","created_at":"2026-07-05T03:14:59.926426+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.07684v1","created_at":"2026-07-05T03:14:59.926426+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.07684","created_at":"2026-07-05T03:14:59.926426+00:00"},{"alias_kind":"pith_short_12","alias_value":"MM67ROAGQZL4","created_at":"2026-07-05T03:14:59.926426+00:00"},{"alias_kind":"pith_short_16","alias_value":"MM67ROAGQZL4RVLW","created_at":"2026-07-05T03:14:59.926426+00:00"},{"alias_kind":"pith_short_8","alias_value":"MM67ROAG","created_at":"2026-07-05T03:14:59.926426+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2112.04359","citing_title":"Ethical and social risks of harm from Language Models","ref_index":293,"is_internal_anchor":false},{"citing_arxiv_id":"2205.01068","citing_title":"OPT: Open Pre-trained Transformer Language Models","ref_index":200,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3","json":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3.json","graph_json":"https://pith.science/api/pith-number/MM67ROAGQZL4RVLWRCE5IHMZL3/graph.json","events_json":"https://pith.science/api/pith-number/MM67ROAGQZL4RVLWRCE5IHMZL3/events.json","paper":"https://pith.science/paper/MM67ROAG"},"agent_actions":{"view_html":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3","download_json":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3.json","view_paper":"https://pith.science/paper/MM67ROAG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.07684&json=true","fetch_graph":"https://pith.science/api/pith-number/MM67ROAGQZL4RVLWRCE5IHMZL3/graph.json","fetch_events":"https://pith.science/api/pith-number/MM67ROAGQZL4RVLWRCE5IHMZL3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3/action/storage_attestation","attest_author":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3/action/author_attestation","sign_citation":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3/action/citation_signature","submit_replication":"https://pith.science/pith/MM67ROAGQZL4RVLWRCE5IHMZL3/action/replication_record"}},"created_at":"2026-07-05T03:14:59.926426+00:00","updated_at":"2026-07-05T03:14:59.926426+00:00"}