{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:DSWBWUNPLXWRZXRUROOALM5UUZ","short_pith_number":"pith:DSWBWUNP","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"},"canonical_sha256":"1cac1b51af5ded1cde348b9c05b3b4a6788701a15957ec743aff4a1058340e5a","source":{"kind":"arxiv","id":"2305.03937","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.03937","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"arxiv_version","alias_value":"2305.03937v1","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.03937","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"pith_short_12","alias_value":"DSWBWUNPLXWR","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"pith_short_16","alias_value":"DSWBWUNPLXWRZXRU","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"pith_short_8","alias_value":"DSWBWUNP","created_at":"2026-07-05T06:07:42Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:DSWBWUNPLXWRZXRUROOALM5UUZ","target":"record","payload":{"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"},"canonical_sha256":"1cac1b51af5ded1cde348b9c05b3b4a6788701a15957ec743aff4a1058340e5a","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"},"source_kind":"arxiv","source_id":"2305.03937","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T06:07:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ccw/S1by1s/3/QZ2x9+xaUzqfiMwd5OLK95iONtCsfQo9QSIpOezOHzkpih796jtAVDtys5bnEKvmmYoTwuzBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-11T05:33:39.944046Z"},"content_sha256":"f6872bfef0395f24bbbab7378d64d4f4482893df7d0a35a7f21b6c0328a90ce9","schema_version":"1.0","event_id":"sha256:f6872bfef0395f24bbbab7378d64d4f4482893df7d0a35a7f21b6c0328a90ce9"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:DSWBWUNPLXWRZXRUROOALM5UUZ","target":"graph","payload":{"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T06:07:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pQ018VwaRwnSmUbXZzHTQF69Lu421+xd3Q4vE4F5/LwVCM6B7XBjlSXDTUp7oNc88h++DjRl+nFfQ2rGfJZTDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-11T05:33:39.944430Z"},"content_sha256":"12c7dd8bce5faa883d92d5722ab0d8b6fabdc00b385d18e4e1ec95563e24429f","schema_version":"1.0","event_id":"sha256:12c7dd8bce5faa883d92d5722ab0d8b6fabdc00b385d18e4e1ec95563e24429f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/bundle.json","state_url":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-11T05:33:39Z","links":{"resolver":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ","bundle":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/bundle.json","state":"https://pith.science/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DSWBWUNPLXWRZXRUROOALM5UUZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:DSWBWUNPLXWRZXRUROOALM5UUZ","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"41515bbd24641e9eba58d10609769acf07fc328a2199d209eb8f2e91cafbf820","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-06T05:35:14Z","title_canon_sha256":"4956a5327695c8a6184cf76fca3cba9a7aa62551e72380433f42cadb9cc33acb"},"schema_version":"1.0","source":{"id":"2305.03937","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.03937","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"arxiv_version","alias_value":"2305.03937v1","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.03937","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"pith_short_12","alias_value":"DSWBWUNPLXWR","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"pith_short_16","alias_value":"DSWBWUNPLXWRZXRU","created_at":"2026-07-05T06:07:42Z"},{"alias_kind":"pith_short_8","alias_value":"DSWBWUNP","created_at":"2026-07-05T06:07:42Z"}],"graph_snapshots":[{"event_id":"sha256:12c7dd8bce5faa883d92d5722ab0d8b6fabdc00b385d18e4e1ec95563e24429f","target":"graph","created_at":"2026-07-05T06:07:42Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2305.03937/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"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","authors_text":"Amjad Almahairi, Anastasia Razdaibiedina, Jimmy Ba, Madian Khabsa, Mike Lewis, Rui Hou, Yuning Mao","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-06T05:35:14Z","title":"Residual Prompt Tuning: Improving Prompt Tuning with Residual Reparameterization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.03937","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f6872bfef0395f24bbbab7378d64d4f4482893df7d0a35a7f21b6c0328a90ce9","target":"record","created_at":"2026-07-05T06:07:42Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"41515bbd24641e9eba58d10609769acf07fc328a2199d209eb8f2e91cafbf820","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-06T05:35:14Z","title_canon_sha256":"4956a5327695c8a6184cf76fca3cba9a7aa62551e72380433f42cadb9cc33acb"},"schema_version":"1.0","source":{"id":"2305.03937","kind":"arxiv","version":1}},"canonical_sha256":"1cac1b51af5ded1cde348b9c05b3b4a6788701a15957ec743aff4a1058340e5a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1cac1b51af5ded1cde348b9c05b3b4a6788701a15957ec743aff4a1058340e5a","first_computed_at":"2026-07-05T06:07:42.861424Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:07:42.861424Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"v6GDR9oBD4X2BHN4eCLa1ChCQU+e+Y870EKwVWuMzcMlBM2B/H76Sl0UoVPILK95fNIdyH8Ir1FJKRI3F1oDCg==","signature_status":"signed_v1","signed_at":"2026-07-05T06:07:42.861819Z","signed_message":"canonical_sha256_bytes"},"source_id":"2305.03937","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f6872bfef0395f24bbbab7378d64d4f4482893df7d0a35a7f21b6c0328a90ce9","sha256:12c7dd8bce5faa883d92d5722ab0d8b6fabdc00b385d18e4e1ec95563e24429f"],"state_sha256":"1a95cb7a852758d5db612635f266c70298afca5333df93c6f78d23b3245797d4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ENzPbaxo9qtmAGfLDSiaa4dkJVBsUfdtm53cWhZLh6l+dNzYGNOmpYyGWC4V6AUA/lWg1urhfI6Y2YH6a6UyDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-11T05:33:39.946592Z","bundle_sha256":"692af229ae3c0443774ac7d8baad4d56d998c20f0b939772313ecb2ed8bbbe4c"}}