{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:LDZRMOBQW5J7S67WOITWOV6WGT","short_pith_number":"pith:LDZRMOBQ","schema_version":"1.0","canonical_sha256":"58f3163830b753f97bf672276757d634fb250635d566fdb310e76d0c60e8d6d8","source":{"kind":"arxiv","id":"2308.15982","version":1},"attestation_state":"computed","paper":{"title":"MerA: Merging Pretrained Adapters For Few-Shot Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dacheng Tao, Liang Ding, Li Shen, Run-Ze Fan, Shwai He, Tianyi Zhou","submitted_at":"2023-08-30T12:10:17Z","abstract_excerpt":"Adapter tuning, which updates only a few parameters, has become a mainstream method for fine-tuning pretrained language models to downstream tasks. However, it often yields subpar results in few-shot learning. AdapterFusion, which assembles pretrained adapters using composition layers tailored to specific tasks, is a possible solution but significantly increases trainable parameters and deployment costs. Despite this, our preliminary study reveals that even single adapters can outperform Adapterfusion in few-shot learning, urging us to propose \\textbf{\\texttt{Merging Pretrained Adapters}} (Mer"},"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":"2308.15982","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-08-30T12:10:17Z","cross_cats_sorted":[],"title_canon_sha256":"76783f5dc40df4b2c254fb3be42e69f46b0392f00cd7b65a23963db168ed2caf","abstract_canon_sha256":"5b8fc4f560d4f12a9696f27ae001c4d7c5361306a50599e1b80b491799185233"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:46:16.399927Z","signature_b64":"lgoQmChCaQI7TEqxpz1RbUAlikLfAkfPCU2swm4omAUjiK/VLI4GbB8jkD6xrIVmrg9N9CtoiaW+odu655hBCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"58f3163830b753f97bf672276757d634fb250635d566fdb310e76d0c60e8d6d8","last_reissued_at":"2026-07-05T06:46:16.399442Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:46:16.399442Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MerA: Merging Pretrained Adapters For Few-Shot Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dacheng Tao, Liang Ding, Li Shen, Run-Ze Fan, Shwai He, Tianyi Zhou","submitted_at":"2023-08-30T12:10:17Z","abstract_excerpt":"Adapter tuning, which updates only a few parameters, has become a mainstream method for fine-tuning pretrained language models to downstream tasks. However, it often yields subpar results in few-shot learning. AdapterFusion, which assembles pretrained adapters using composition layers tailored to specific tasks, is a possible solution but significantly increases trainable parameters and deployment costs. Despite this, our preliminary study reveals that even single adapters can outperform Adapterfusion in few-shot learning, urging us to propose \\textbf{\\texttt{Merging Pretrained Adapters}} (Mer"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2308.15982","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/2308.15982/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":"2308.15982","created_at":"2026-07-05T06:46:16.399501+00:00"},{"alias_kind":"arxiv_version","alias_value":"2308.15982v1","created_at":"2026-07-05T06:46:16.399501+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.15982","created_at":"2026-07-05T06:46:16.399501+00:00"},{"alias_kind":"pith_short_12","alias_value":"LDZRMOBQW5J7","created_at":"2026-07-05T06:46:16.399501+00:00"},{"alias_kind":"pith_short_16","alias_value":"LDZRMOBQW5J7S67W","created_at":"2026-07-05T06:46:16.399501+00:00"},{"alias_kind":"pith_short_8","alias_value":"LDZRMOBQ","created_at":"2026-07-05T06:46:16.399501+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2408.07666","citing_title":"Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2403.14608","citing_title":"Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT","json":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT.json","graph_json":"https://pith.science/api/pith-number/LDZRMOBQW5J7S67WOITWOV6WGT/graph.json","events_json":"https://pith.science/api/pith-number/LDZRMOBQW5J7S67WOITWOV6WGT/events.json","paper":"https://pith.science/paper/LDZRMOBQ"},"agent_actions":{"view_html":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT","download_json":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT.json","view_paper":"https://pith.science/paper/LDZRMOBQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2308.15982&json=true","fetch_graph":"https://pith.science/api/pith-number/LDZRMOBQW5J7S67WOITWOV6WGT/graph.json","fetch_events":"https://pith.science/api/pith-number/LDZRMOBQW5J7S67WOITWOV6WGT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT/action/storage_attestation","attest_author":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT/action/author_attestation","sign_citation":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT/action/citation_signature","submit_replication":"https://pith.science/pith/LDZRMOBQW5J7S67WOITWOV6WGT/action/replication_record"}},"created_at":"2026-07-05T06:46:16.399501+00:00","updated_at":"2026-07-05T06:46:16.399501+00:00"}