{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:B5S46HPMLRYCWZOQI37L2WDPIM","short_pith_number":"pith:B5S46HPM","schema_version":"1.0","canonical_sha256":"0f65cf1dec5c702b65d046febd586f4314ec674a9bb3e3885e86f7312a57da36","source":{"kind":"arxiv","id":"2505.02130","version":1},"attestation_state":"computed","paper":{"title":"Attention Mechanisms Perspective: Exploring LLM Processing of Graph-Structured Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Hongke Zhao, Jianpin Fan, Likang Wu, Ming He, Zhong Guan","submitted_at":"2025-05-04T14:40:31Z","abstract_excerpt":"Attention mechanisms are critical to the success of large language models (LLMs), driving significant advancements in multiple fields. However, for graph-structured data, which requires emphasis on topological connections, they fall short compared to message-passing mechanisms on fixed links, such as those employed by Graph Neural Networks (GNNs). This raises a question: ``Does attention fail for graphs in natural language settings?'' Motivated by these observations, we embarked on an empirical study from the perspective of attention mechanisms to explore how LLMs process graph-structured data"},"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":"2505.02130","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-04T14:40:31Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"19e27983b8df679a2880680b4066d8ed2f79303f3945aa7ae8c19940dd9b17a4","abstract_canon_sha256":"e2289fa8568b1fe1edc75a284836b265f7c59336df4f91d41e19507986d83e65"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:58:32.438927Z","signature_b64":"hQckQgBFpapEgvZSRss47AJq5/oAwe/KuT2tl9jHF4Ukdof+xSi2Dy4jAbFOBD6CSKjk7Ac39Q0vgGCri4R8DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0f65cf1dec5c702b65d046febd586f4314ec674a9bb3e3885e86f7312a57da36","last_reissued_at":"2026-07-05T10:58:32.438375Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:58:32.438375Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Attention Mechanisms Perspective: Exploring LLM Processing of Graph-Structured Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Hongke Zhao, Jianpin Fan, Likang Wu, Ming He, Zhong Guan","submitted_at":"2025-05-04T14:40:31Z","abstract_excerpt":"Attention mechanisms are critical to the success of large language models (LLMs), driving significant advancements in multiple fields. However, for graph-structured data, which requires emphasis on topological connections, they fall short compared to message-passing mechanisms on fixed links, such as those employed by Graph Neural Networks (GNNs). This raises a question: ``Does attention fail for graphs in natural language settings?'' Motivated by these observations, we embarked on an empirical study from the perspective of attention mechanisms to explore how LLMs process graph-structured data"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.02130","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/2505.02130/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":"2505.02130","created_at":"2026-07-05T10:58:32.438463+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.02130v1","created_at":"2026-07-05T10:58:32.438463+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.02130","created_at":"2026-07-05T10:58:32.438463+00:00"},{"alias_kind":"pith_short_12","alias_value":"B5S46HPMLRYC","created_at":"2026-07-05T10:58:32.438463+00:00"},{"alias_kind":"pith_short_16","alias_value":"B5S46HPMLRYCWZOQ","created_at":"2026-07-05T10:58:32.438463+00:00"},{"alias_kind":"pith_short_8","alias_value":"B5S46HPM","created_at":"2026-07-05T10:58:32.438463+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03712","citing_title":"When Graph Tokens Sink: A Mechanistic Analysis of Graph Language Models","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19028","citing_title":"Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors","ref_index":298,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM","json":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM.json","graph_json":"https://pith.science/api/pith-number/B5S46HPMLRYCWZOQI37L2WDPIM/graph.json","events_json":"https://pith.science/api/pith-number/B5S46HPMLRYCWZOQI37L2WDPIM/events.json","paper":"https://pith.science/paper/B5S46HPM"},"agent_actions":{"view_html":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM","download_json":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM.json","view_paper":"https://pith.science/paper/B5S46HPM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.02130&json=true","fetch_graph":"https://pith.science/api/pith-number/B5S46HPMLRYCWZOQI37L2WDPIM/graph.json","fetch_events":"https://pith.science/api/pith-number/B5S46HPMLRYCWZOQI37L2WDPIM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM/action/storage_attestation","attest_author":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM/action/author_attestation","sign_citation":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM/action/citation_signature","submit_replication":"https://pith.science/pith/B5S46HPMLRYCWZOQI37L2WDPIM/action/replication_record"}},"created_at":"2026-07-05T10:58:32.438463+00:00","updated_at":"2026-07-05T10:58:32.438463+00:00"}