{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:74K5VCBV6M5IT6U43EGTNRSMQA","short_pith_number":"pith:74K5VCBV","schema_version":"1.0","canonical_sha256":"ff15da8835f33a89fa9cd90d36c64c80109dfc0c8217f619604a7abbc88d55e0","source":{"kind":"arxiv","id":"2211.08099","version":2},"attestation_state":"computed","paper":{"title":"A Universal Discriminator for Zero-Shot Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Haike Xu, Jing Zhou, Yanan Zheng, Zhilin Yang, Zongyu Lin","submitted_at":"2022-11-15T12:33:31Z","abstract_excerpt":"Generative modeling has been the dominant approach for large-scale pretraining and zero-shot generalization. In this work, we challenge this convention by showing that discriminative approaches perform substantially better than generative ones on a large number of NLP tasks. Technically, we train a single discriminator to predict whether a text sample comes from the true data distribution, similar to GANs. Since many NLP tasks can be formulated as selecting from a few options, we use this discriminator to predict the concatenation of input and which option has the highest probability of coming"},"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":"2211.08099","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-11-15T12:33:31Z","cross_cats_sorted":[],"title_canon_sha256":"dfe7fe2af98670a385e002b7f01d4bd4ba79f00816e9c6b8b7724cb3c9e02630","abstract_canon_sha256":"a25f7cc036babce5f36c1c291075453557854404fe6ff21d25d76e10578def51"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:17:42.335465Z","signature_b64":"RHffFjcS2mmXCnJgrqkfw2PYGJUqr87f+3TIX9Et2oTFeNMsQUzVVMFffnwf+fjrS1z1O5JoirR2DtZQAmoOAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff15da8835f33a89fa9cd90d36c64c80109dfc0c8217f619604a7abbc88d55e0","last_reissued_at":"2026-07-05T06:17:42.334979Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:17:42.334979Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Universal Discriminator for Zero-Shot Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Haike Xu, Jing Zhou, Yanan Zheng, Zhilin Yang, Zongyu Lin","submitted_at":"2022-11-15T12:33:31Z","abstract_excerpt":"Generative modeling has been the dominant approach for large-scale pretraining and zero-shot generalization. In this work, we challenge this convention by showing that discriminative approaches perform substantially better than generative ones on a large number of NLP tasks. Technically, we train a single discriminator to predict whether a text sample comes from the true data distribution, similar to GANs. Since many NLP tasks can be formulated as selecting from a few options, we use this discriminator to predict the concatenation of input and which option has the highest probability of coming"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.08099","kind":"arxiv","version":2},"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/2211.08099/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":"2211.08099","created_at":"2026-07-05T06:17:42.335036+00:00"},{"alias_kind":"arxiv_version","alias_value":"2211.08099v2","created_at":"2026-07-05T06:17:42.335036+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.08099","created_at":"2026-07-05T06:17:42.335036+00:00"},{"alias_kind":"pith_short_12","alias_value":"74K5VCBV6M5I","created_at":"2026-07-05T06:17:42.335036+00:00"},{"alias_kind":"pith_short_16","alias_value":"74K5VCBV6M5IT6U4","created_at":"2026-07-05T06:17:42.335036+00:00"},{"alias_kind":"pith_short_8","alias_value":"74K5VCBV","created_at":"2026-07-05T06:17:42.335036+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.21848","citing_title":"FPAN: Mitigating Replication in Diffusion Models through the Fine-Grained Probabilistic Addition of Noise to Token Embeddings","ref_index":33,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA","json":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA.json","graph_json":"https://pith.science/api/pith-number/74K5VCBV6M5IT6U43EGTNRSMQA/graph.json","events_json":"https://pith.science/api/pith-number/74K5VCBV6M5IT6U43EGTNRSMQA/events.json","paper":"https://pith.science/paper/74K5VCBV"},"agent_actions":{"view_html":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA","download_json":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA.json","view_paper":"https://pith.science/paper/74K5VCBV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2211.08099&json=true","fetch_graph":"https://pith.science/api/pith-number/74K5VCBV6M5IT6U43EGTNRSMQA/graph.json","fetch_events":"https://pith.science/api/pith-number/74K5VCBV6M5IT6U43EGTNRSMQA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA/action/storage_attestation","attest_author":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA/action/author_attestation","sign_citation":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA/action/citation_signature","submit_replication":"https://pith.science/pith/74K5VCBV6M5IT6U43EGTNRSMQA/action/replication_record"}},"created_at":"2026-07-05T06:17:42.335036+00:00","updated_at":"2026-07-05T06:17:42.335036+00:00"}