{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:6XBLTDCH77HVIH2NDPG3UJQNT2","short_pith_number":"pith:6XBLTDCH","schema_version":"1.0","canonical_sha256":"f5c2b98c47ffcf541f4d1bcdba260d9ea34295767ae3af1465612e95e8e70cb8","source":{"kind":"arxiv","id":"2010.11934","version":3},"attestation_state":"computed","paper":{"title":"mT5: A massively multilingual pre-trained text-to-text transformer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adam Roberts, Aditya Barua, Aditya Siddhant, Colin Raffel, Linting Xue, Mihir Kale, Noah Constant, Rami Al-Rfou","submitted_at":"2020-10-22T17:58:14Z","abstract_excerpt":"The recent \"Text-to-Text Transfer Transformer\" (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 that was pre-trained on a new Common Crawl-based dataset covering 101 languages. We detail the design and modified training of mT5 and demonstrate its state-of-the-art performance on many multilingual benchmarks. We also describe a simple technique to prevent \"accidental translation\" in the zero-shot setting, where a generative model chooses to (part"},"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":"2010.11934","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-22T17:58:14Z","cross_cats_sorted":[],"title_canon_sha256":"0543b28134c097d7464d5bca55551922817efb8417c6b766938362bd725cfdf7","abstract_canon_sha256":"3294fe684013b82b5768b76e246048fca843970c418abc521f549e0e92a44c6d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:22:16.956683Z","signature_b64":"KmgUfIn7fkSpfjiuO2n00MU6XwaNSooswVgVxAqHJRK9Njejj9JgXPJmJJ9pXOju9uquVyaCD0zHX6qwRKHMBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f5c2b98c47ffcf541f4d1bcdba260d9ea34295767ae3af1465612e95e8e70cb8","last_reissued_at":"2026-07-05T02:22:16.956205Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:22:16.956205Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"mT5: A massively multilingual pre-trained text-to-text transformer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adam Roberts, Aditya Barua, Aditya Siddhant, Colin Raffel, Linting Xue, Mihir Kale, Noah Constant, Rami Al-Rfou","submitted_at":"2020-10-22T17:58:14Z","abstract_excerpt":"The recent \"Text-to-Text Transfer Transformer\" (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 that was pre-trained on a new Common Crawl-based dataset covering 101 languages. We detail the design and modified training of mT5 and demonstrate its state-of-the-art performance on many multilingual benchmarks. We also describe a simple technique to prevent \"accidental translation\" in the zero-shot setting, where a generative model chooses to (part"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.11934","kind":"arxiv","version":3},"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/2010.11934/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":"2010.11934","created_at":"2026-07-05T02:22:16.956273+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.11934v3","created_at":"2026-07-05T02:22:16.956273+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.11934","created_at":"2026-07-05T02:22:16.956273+00:00"},{"alias_kind":"pith_short_12","alias_value":"6XBLTDCH77HV","created_at":"2026-07-05T02:22:16.956273+00:00"},{"alias_kind":"pith_short_16","alias_value":"6XBLTDCH77HVIH2N","created_at":"2026-07-05T02:22:16.956273+00:00"},{"alias_kind":"pith_short_8","alias_value":"6XBLTDCH","created_at":"2026-07-05T02:22:16.956273+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":27,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06544","citing_title":"Rethinking Indic AI from a Lens of Cultural Heritage Preservation","ref_index":240,"is_internal_anchor":true},{"citing_arxiv_id":"2606.20993","citing_title":"Phonemes to the Rescue: Multilingual Tokenization Based on International Phonetic Alphabet","ref_index":65,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09767","citing_title":"Data Synthesis and Parameter-Efficient Fine-Tuning for Low-Resource NMT: A Case Study on Q'eqchi' Mayan","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06738","citing_title":"Modular Monolingual Adaptation using Pretrained Language Models","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04661","citing_title":"CRAFT: Cost-aware Refinement And Front-aware Tuning of Prompts","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2107.06499","citing_title":"Deduplicating Training Data Makes Language Models Better","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2204.06745","citing_title":"GPT-NeoX-20B: An Open-Source Autoregressive Language Model","ref_index":105,"is_internal_anchor":false},{"citing_arxiv_id":"2407.06048","citing_title":"Vision-Braille: A Curriculum Learning Toolkit and Braille-Chinese Corpus for Braille Translation","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2412.00131","citing_title":"Open-Sora Plan: Open-Source Large Video Generation Model","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17152","citing_title":"Multilingual and Multimodal LLMs in the Wild: Building for Low-Resource Languages","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2307.06435","citing_title":"A Comprehensive Overview of Large Language Models","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2508.17458","citing_title":"Evaluating the Impact of Verbal Multiword Expressions on Machine Translation","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2305.16264","citing_title":"Scaling Data-Constrained Language Models","ref_index":134,"is_internal_anchor":false},{"citing_arxiv_id":"2311.16867","citing_title":"The Falcon Series of Open Language Models","ref_index":83,"is_internal_anchor":false},{"citing_arxiv_id":"2101.03961","citing_title":"Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2202.08906","citing_title":"ST-MoE: Designing Stable and Transferable Sparse Expert Models","ref_index":209,"is_internal_anchor":false},{"citing_arxiv_id":"2402.03216","citing_title":"M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23779","citing_title":"GLIER: Generative Legal Inference and Evidence Ranking for Legal Case Retrieval","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00618","citing_title":"Is Textual Similarity Invariant under Machine Translation? Evidence Based on the Political Manifesto Corpus","ref_index":77,"is_internal_anchor":false},{"citing_arxiv_id":"2402.06196","citing_title":"Large Language Models: A Survey","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19508","citing_title":"Bangla Key2Text: Text Generation from Keywords for a Low Resource Language","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2407.07726","citing_title":"PaliGemma: A versatile 3B VLM for transfer","ref_index":152,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10590","citing_title":"Bridging Linguistic Gaps: Cross-Lingual Mapping in Pre-Training and Dataset for Enhanced Multilingual LLM Performance","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10417","citing_title":"LASQ: A Low-resource Aspect-based Sentiment Quadruple Extraction Dataset","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19797","citing_title":"Enhancing ASR Performance in the Medical Domain for Dravidian Languages","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2","json":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2.json","graph_json":"https://pith.science/api/pith-number/6XBLTDCH77HVIH2NDPG3UJQNT2/graph.json","events_json":"https://pith.science/api/pith-number/6XBLTDCH77HVIH2NDPG3UJQNT2/events.json","paper":"https://pith.science/paper/6XBLTDCH"},"agent_actions":{"view_html":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2","download_json":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2.json","view_paper":"https://pith.science/paper/6XBLTDCH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.11934&json=true","fetch_graph":"https://pith.science/api/pith-number/6XBLTDCH77HVIH2NDPG3UJQNT2/graph.json","fetch_events":"https://pith.science/api/pith-number/6XBLTDCH77HVIH2NDPG3UJQNT2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2/action/storage_attestation","attest_author":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2/action/author_attestation","sign_citation":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2/action/citation_signature","submit_replication":"https://pith.science/pith/6XBLTDCH77HVIH2NDPG3UJQNT2/action/replication_record"}},"created_at":"2026-07-05T02:22:16.956273+00:00","updated_at":"2026-07-05T02:22:16.956273+00:00"}