{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PEYCWU4UIY62X2B54OQZCB2U64","short_pith_number":"pith:PEYCWU4U","schema_version":"1.0","canonical_sha256":"79302b5394463dabe83de3a1910754f72c26f461f8425806269356e6c0d127c3","source":{"kind":"arxiv","id":"2408.01119","version":4},"attestation_state":"computed","paper":{"title":"Task Prompt Vectors: Effective Initialization through Multi-Task Soft-Prompt Transfer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ivan Srba, Maria Bielikova, Robert Belanec, Simon Ostermann","submitted_at":"2024-08-02T09:00:03Z","abstract_excerpt":"Prompt tuning is an efficient solution for training large language models (LLMs). However, current soft-prompt-based methods often sacrifice multi-task modularity, requiring the training process to be fully or partially repeated for each newly added task. While recent work on task vectors applied arithmetic operations on full model weights to achieve the desired multi-task performance, a similar approach for soft-prompts is still missing. To this end, we introduce Task Prompt Vectors, created by element-wise difference between weights of tuned soft-prompts and their random initialization. Expe"},"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":"2408.01119","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-08-02T09:00:03Z","cross_cats_sorted":[],"title_canon_sha256":"579246b4421dfd4b9f3df20c2a70651ec7e7b8a56038586e772998499670329b","abstract_canon_sha256":"4c088bae37bf9563031668dd7299709558a8ca234b6f7593b5d13484ccbc4b3c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:10:12.990645Z","signature_b64":"4BhLvhdLbhtZAOcfuco8GbfhvZr/LIeQMckPwoDzFa6B2VCkVZS6/6AOIo/b98VqTL25US8V44BHiGftm7xaDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"79302b5394463dabe83de3a1910754f72c26f461f8425806269356e6c0d127c3","last_reissued_at":"2026-05-18T03:10:12.989855Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:10:12.989855Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Task Prompt Vectors: Effective Initialization through Multi-Task Soft-Prompt Transfer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ivan Srba, Maria Bielikova, Robert Belanec, Simon Ostermann","submitted_at":"2024-08-02T09:00:03Z","abstract_excerpt":"Prompt tuning is an efficient solution for training large language models (LLMs). However, current soft-prompt-based methods often sacrifice multi-task modularity, requiring the training process to be fully or partially repeated for each newly added task. While recent work on task vectors applied arithmetic operations on full model weights to achieve the desired multi-task performance, a similar approach for soft-prompts is still missing. To this end, we introduce Task Prompt Vectors, created by element-wise difference between weights of tuned soft-prompts and their random initialization. Expe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.01119","kind":"arxiv","version":4},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"2408.01119","created_at":"2026-05-18T03:10:12.989967+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.01119v4","created_at":"2026-05-18T03:10:12.989967+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.01119","created_at":"2026-05-18T03:10:12.989967+00:00"},{"alias_kind":"pith_short_12","alias_value":"PEYCWU4UIY62","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"PEYCWU4UIY62X2B5","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"PEYCWU4U","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2408.07666","citing_title":"Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64","json":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64.json","graph_json":"https://pith.science/api/pith-number/PEYCWU4UIY62X2B54OQZCB2U64/graph.json","events_json":"https://pith.science/api/pith-number/PEYCWU4UIY62X2B54OQZCB2U64/events.json","paper":"https://pith.science/paper/PEYCWU4U"},"agent_actions":{"view_html":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64","download_json":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64.json","view_paper":"https://pith.science/paper/PEYCWU4U","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.01119&json=true","fetch_graph":"https://pith.science/api/pith-number/PEYCWU4UIY62X2B54OQZCB2U64/graph.json","fetch_events":"https://pith.science/api/pith-number/PEYCWU4UIY62X2B54OQZCB2U64/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64/action/storage_attestation","attest_author":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64/action/author_attestation","sign_citation":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64/action/citation_signature","submit_replication":"https://pith.science/pith/PEYCWU4UIY62X2B54OQZCB2U64/action/replication_record"}},"created_at":"2026-05-18T03:10:12.989967+00:00","updated_at":"2026-05-18T03:10:12.989967+00:00"}