{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:VGJSO6UMCQX5D6RKNYTZTESYNN","short_pith_number":"pith:VGJSO6UM","schema_version":"1.0","canonical_sha256":"a993277a8c142fd1fa2a6e279992586b534fc00140b59e2c1df94d579581c976","source":{"kind":"arxiv","id":"2109.03630","version":1},"attestation_state":"computed","paper":{"title":"Discrete and Soft Prompting for Multilingual Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hinrich Sch\\\"utze, Mengjie Zhao","submitted_at":"2021-09-08T13:22:59Z","abstract_excerpt":"It has been shown for English that discrete and soft prompting perform strongly in few-shot learning with pretrained language models (PLMs). In this paper, we show that discrete and soft prompting perform better than finetuning in multilingual cases: Crosslingual transfer and in-language training of multilingual natural language inference. For example, with 48 English training examples, finetuning obtains 33.74% accuracy in crosslingual transfer, barely surpassing the majority baseline (33.33%). In contrast, discrete and soft prompting outperform finetuning, achieving 36.43% and 38.79%. We als"},"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":"2109.03630","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-09-08T13:22:59Z","cross_cats_sorted":[],"title_canon_sha256":"faf01af8cff38db22a30b3adbd5d747275d307d138abd2ffa731706d13d519e7","abstract_canon_sha256":"46285e5b73a01858d2582781b4d16086b72a250a39e440226aec97d3c4b71be0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:12:41.390299Z","signature_b64":"vDrjXQ2Mwf2ppTT5Bjea681S+Y5KqktVOdHV1+PgNYC09Fn1k0BAlidlpTs24XYAGSKHgXv5yNN6N/8ySKJ3Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a993277a8c142fd1fa2a6e279992586b534fc00140b59e2c1df94d579581c976","last_reissued_at":"2026-07-05T03:12:41.389783Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:12:41.389783Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Discrete and Soft Prompting for Multilingual Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hinrich Sch\\\"utze, Mengjie Zhao","submitted_at":"2021-09-08T13:22:59Z","abstract_excerpt":"It has been shown for English that discrete and soft prompting perform strongly in few-shot learning with pretrained language models (PLMs). In this paper, we show that discrete and soft prompting perform better than finetuning in multilingual cases: Crosslingual transfer and in-language training of multilingual natural language inference. For example, with 48 English training examples, finetuning obtains 33.74% accuracy in crosslingual transfer, barely surpassing the majority baseline (33.33%). In contrast, discrete and soft prompting outperform finetuning, achieving 36.43% and 38.79%. We als"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.03630","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/2109.03630/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":"2109.03630","created_at":"2026-07-05T03:12:41.389844+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.03630v1","created_at":"2026-07-05T03:12:41.389844+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.03630","created_at":"2026-07-05T03:12:41.389844+00:00"},{"alias_kind":"pith_short_12","alias_value":"VGJSO6UMCQX5","created_at":"2026-07-05T03:12:41.389844+00:00"},{"alias_kind":"pith_short_16","alias_value":"VGJSO6UMCQX5D6RK","created_at":"2026-07-05T03:12:41.389844+00:00"},{"alias_kind":"pith_short_8","alias_value":"VGJSO6UM","created_at":"2026-07-05T03:12:41.389844+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2205.01068","citing_title":"OPT: Open Pre-trained Transformer Language Models","ref_index":190,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN","json":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN.json","graph_json":"https://pith.science/api/pith-number/VGJSO6UMCQX5D6RKNYTZTESYNN/graph.json","events_json":"https://pith.science/api/pith-number/VGJSO6UMCQX5D6RKNYTZTESYNN/events.json","paper":"https://pith.science/paper/VGJSO6UM"},"agent_actions":{"view_html":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN","download_json":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN.json","view_paper":"https://pith.science/paper/VGJSO6UM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.03630&json=true","fetch_graph":"https://pith.science/api/pith-number/VGJSO6UMCQX5D6RKNYTZTESYNN/graph.json","fetch_events":"https://pith.science/api/pith-number/VGJSO6UMCQX5D6RKNYTZTESYNN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN/action/storage_attestation","attest_author":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN/action/author_attestation","sign_citation":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN/action/citation_signature","submit_replication":"https://pith.science/pith/VGJSO6UMCQX5D6RKNYTZTESYNN/action/replication_record"}},"created_at":"2026-07-05T03:12:41.389844+00:00","updated_at":"2026-07-05T03:12:41.389844+00:00"}