{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:A4464TP72GKGPIK7GKUNTXID6S","short_pith_number":"pith:A4464TP7","schema_version":"1.0","canonical_sha256":"0739ee4dffd19467a15f32a8d9dd03f4b992ab4898f484413f0825fabaf41d71","source":{"kind":"arxiv","id":"2402.11709","version":2},"attestation_state":"computed","paper":{"title":"GNNavi: Navigating the Information Flow in Large Language Models by Graph Neural Network","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ercong Nie, Helmut Schmid, Hinrich Sch\\\"utze, Michael F\\\"arber, Shuzhou Yuan","submitted_at":"2024-02-18T21:13:05Z","abstract_excerpt":"Large Language Models (LLMs) exhibit strong In-Context Learning (ICL) capabilities when prompts with demonstrations are used. However, fine-tuning still remains crucial to further enhance their adaptability. Prompt-based fine-tuning proves to be an effective fine-tuning method in low-data scenarios, but high demands on computing resources limit its practicality. We address this issue by introducing a prompt-based parameter-efficient fine-tuning (PEFT) approach. GNNavi leverages insights into ICL's information flow dynamics, which indicates that label words act in prompts as anchors for informa"},"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":"2402.11709","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-18T21:13:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d18d6df1dbeafb432c9f4833b419b33aefe0d35526b3c6f304ba2a6b188edc34","abstract_canon_sha256":"1d4e85d0151194b2f1def2af9e5421bc132e04cf702afdfc8158b4a45ec67286"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:38.485033Z","signature_b64":"K9ZRf9ioNBrHJ8Ey08N2CHIRzKmRZKGmqDt1hFX1ePaJq/xHE5B4nyGGkjKNtR7gAuLDkjFWngS4Xwq/KNidAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0739ee4dffd19467a15f32a8d9dd03f4b992ab4898f484413f0825fabaf41d71","last_reissued_at":"2026-07-05T08:28:38.484580Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:38.484580Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"GNNavi: Navigating the Information Flow in Large Language Models by Graph Neural Network","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ercong Nie, Helmut Schmid, Hinrich Sch\\\"utze, Michael F\\\"arber, Shuzhou Yuan","submitted_at":"2024-02-18T21:13:05Z","abstract_excerpt":"Large Language Models (LLMs) exhibit strong In-Context Learning (ICL) capabilities when prompts with demonstrations are used. However, fine-tuning still remains crucial to further enhance their adaptability. Prompt-based fine-tuning proves to be an effective fine-tuning method in low-data scenarios, but high demands on computing resources limit its practicality. We address this issue by introducing a prompt-based parameter-efficient fine-tuning (PEFT) approach. GNNavi leverages insights into ICL's information flow dynamics, which indicates that label words act in prompts as anchors for informa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11709","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/2402.11709/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":"2402.11709","created_at":"2026-07-05T08:28:38.484635+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.11709v2","created_at":"2026-07-05T08:28:38.484635+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11709","created_at":"2026-07-05T08:28:38.484635+00:00"},{"alias_kind":"pith_short_12","alias_value":"A4464TP72GKG","created_at":"2026-07-05T08:28:38.484635+00:00"},{"alias_kind":"pith_short_16","alias_value":"A4464TP72GKGPIK7","created_at":"2026-07-05T08:28:38.484635+00:00"},{"alias_kind":"pith_short_8","alias_value":"A4464TP7","created_at":"2026-07-05T08:28:38.484635+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.17458","citing_title":"EHRAG: Bridging Semantic Gaps in Lightweight GraphRAG via Hybrid Hypergraph Construction and Retrieval","ref_index":250,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S","json":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S.json","graph_json":"https://pith.science/api/pith-number/A4464TP72GKGPIK7GKUNTXID6S/graph.json","events_json":"https://pith.science/api/pith-number/A4464TP72GKGPIK7GKUNTXID6S/events.json","paper":"https://pith.science/paper/A4464TP7"},"agent_actions":{"view_html":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S","download_json":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S.json","view_paper":"https://pith.science/paper/A4464TP7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.11709&json=true","fetch_graph":"https://pith.science/api/pith-number/A4464TP72GKGPIK7GKUNTXID6S/graph.json","fetch_events":"https://pith.science/api/pith-number/A4464TP72GKGPIK7GKUNTXID6S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S/action/storage_attestation","attest_author":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S/action/author_attestation","sign_citation":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S/action/citation_signature","submit_replication":"https://pith.science/pith/A4464TP72GKGPIK7GKUNTXID6S/action/replication_record"}},"created_at":"2026-07-05T08:28:38.484635+00:00","updated_at":"2026-07-05T08:28:38.484635+00:00"}