{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:DSWBWUNPLXWRZXRUROOALM5UUZ","short_pith_number":"pith:DSWBWUNP","schema_version":"1.0","canonical_sha256":"1cac1b51af5ded1cde348b9c05b3b4a6788701a15957ec743aff4a1058340e5a","source":{"kind":"arxiv","id":"2305.03937","version":1},"attestation_state":"computed","paper":{"title":"Residual Prompt Tuning: Improving Prompt Tuning with Residual Reparameterization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Amjad Almahairi, Anastasia Razdaibiedina, Jimmy Ba, Madian Khabsa, Mike Lewis, Rui Hou, Yuning Mao","submitted_at":"2023-05-06T05:35:14Z","abstract_excerpt":"Prompt tuning is one of the successful approaches for parameter-efficient tuning of pre-trained language models. Despite being arguably the most parameter-efficient (tuned soft prompts constitute <0.1% of total parameters), it typically performs worse than other efficient tuning methods and is quite sensitive to hyper-parameters. In this work, we introduce Residual Prompt Tuning - a simple and efficient method that significantly improves the performance and stability of prompt tuning. We propose to reparameterize soft prompt embeddings using a shallow network with a residual connection. Our ex"},"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":"2305.03937","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-06T05:35:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4956a5327695c8a6184cf76fca3cba9a7aa62551e72380433f42cadb9cc33acb","abstract_canon_sha256":"41515bbd24641e9eba58d10609769acf07fc328a2199d209eb8f2e91cafbf820"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:07:42.861819Z","signature_b64":"v6GDR9oBD4X2BHN4eCLa1ChCQU+e+Y870EKwVWuMzcMlBM2B/H76Sl0UoVPILK95fNIdyH8Ir1FJKRI3F1oDCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1cac1b51af5ded1cde348b9c05b3b4a6788701a15957ec743aff4a1058340e5a","last_reissued_at":"2026-07-05T06:07:42.861424Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:07:42.861424Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Residual Prompt Tuning: Improving Prompt Tuning with Residual Reparameterization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Amjad Almahairi, Anastasia Razdaibiedina, Jimmy Ba, Madian Khabsa, Mike Lewis, Rui Hou, Yuning Mao","submitted_at":"2023-05-06T05:35:14Z","abstract_excerpt":"Prompt tuning is one of the successful approaches for parameter-efficient tuning of pre-trained language models. Despite being arguably the most parameter-efficient (tuned soft prompts constitute <0.1% of total parameters), it typically performs worse than other efficient tuning methods and is quite sensitive to hyper-parameters. In this work, we introduce Residual Prompt Tuning - a simple and efficient method that significantly improves the performance and stability of prompt tuning. We propose to reparameterize soft prompt embeddings using a shallow network with a residual connection. Our ex"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.03937","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/2305.03937/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":"2305.03937","created_at":"2026-07-05T06:07:42.861480+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.03937v1","created_at":"2026-07-05T06:07:42.861480+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.03937","created_at":"2026-07-05T06:07:42.861480+00:00"},{"alias_kind":"pith_short_12","alias_value":"DSWBWUNPLXWR","created_at":"2026-07-05T06:07:42.861480+00:00"},{"alias_kind":"pith_short_16","alias_value":"DSWBWUNPLXWRZXRU","created_at":"2026-07-05T06:07:42.861480+00:00"},{"alias_kind":"pith_short_8","alias_value":"DSWBWUNP","created_at":"2026-07-05T06:07:42.861480+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.20333","citing_title":"SoftSkill: Behavioral Compression for Contextual Adaptation","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04447","citing_title":"Deep Reprogramming Distillation for Medical Foundation Models","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ","json":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ.json","graph_json":"https://pith.science/api/pith-number/DSWBWUNPLXWRZXRUROOALM5UUZ/graph.json","events_json":"https://pith.science/api/pith-number/DSWBWUNPLXWRZXRUROOALM5UUZ/events.json","paper":"https://pith.science/paper/DSWBWUNP"},"agent_actions":{"view_html":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ","download_json":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ.json","view_paper":"https://pith.science/paper/DSWBWUNP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.03937&json=true","fetch_graph":"https://pith.science/api/pith-number/DSWBWUNPLXWRZXRUROOALM5UUZ/graph.json","fetch_events":"https://pith.science/api/pith-number/DSWBWUNPLXWRZXRUROOALM5UUZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/action/storage_attestation","attest_author":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/action/author_attestation","sign_citation":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/action/citation_signature","submit_replication":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/action/replication_record"}},"created_at":"2026-07-05T06:07:42.861480+00:00","updated_at":"2026-07-05T06:07:42.861480+00:00"}