{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:OY645JZ7E3NMB4NKRN3NKSCNXY","short_pith_number":"pith:OY645JZ7","schema_version":"1.0","canonical_sha256":"763dcea73f26dac0f1aa8b76d5484dbe3cfa02e579e5b64ae23c136d4c86faff","source":{"kind":"arxiv","id":"2602.02518","version":2},"attestation_state":"computed","paper":{"title":"GraphDancer: Training LLMs to Explore and Reason over Graphs via Two-Stage Curriculum Post-Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Jianwen Xie, Ping Nie, Yuyang Bai, Yu Zhang, Zhuofeng Li","submitted_at":"2026-01-24T02:44:49Z","abstract_excerpt":"Large language models (LLMs) increasingly rely on external knowledge to improve factuality, yet many real-world knowledge sources are organized as heterogeneous graphs rather than plain text. Reasoning over such graphs requires models to follow schema-defined relations through precise function calls and to aggregate evidence across multiple rounds of interaction. We propose GraphDancer, a two-stage post-training framework that teaches LLMs to reason over graphs by interleaving natural-language reasoning with graph function execution. The first stage teaches the model how to interact with the g"},"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":"2602.02518","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-24T02:44:49Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"b5ed10c10eaf48faeaef109f41b1a4a35fe1ee587dc920324d80ea79837301e0","abstract_canon_sha256":"8c6265b3c12e1d3a35113ccb79d1e32477e3266e17f20a707452ac96b0903748"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:04:55.852667Z","signature_b64":"3LrdrwAufrSm4vcl+wwW/QCj7eUcHSIbkr/2h5dOmzX/yz/C6zxulPPDx1+8T6Q6xfA/Ar3RpK9LqBCCMC/JBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"763dcea73f26dac0f1aa8b76d5484dbe3cfa02e579e5b64ae23c136d4c86faff","last_reissued_at":"2026-05-27T01:04:55.851848Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:04:55.851848Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"GraphDancer: Training LLMs to Explore and Reason over Graphs via Two-Stage Curriculum Post-Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Jianwen Xie, Ping Nie, Yuyang Bai, Yu Zhang, Zhuofeng Li","submitted_at":"2026-01-24T02:44:49Z","abstract_excerpt":"Large language models (LLMs) increasingly rely on external knowledge to improve factuality, yet many real-world knowledge sources are organized as heterogeneous graphs rather than plain text. Reasoning over such graphs requires models to follow schema-defined relations through precise function calls and to aggregate evidence across multiple rounds of interaction. We propose GraphDancer, a two-stage post-training framework that teaches LLMs to reason over graphs by interleaving natural-language reasoning with graph function execution. The first stage teaches the model how to interact with the g"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.02518","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/2602.02518/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":"2602.02518","created_at":"2026-05-27T01:04:55.851957+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.02518v2","created_at":"2026-05-27T01:04:55.851957+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.02518","created_at":"2026-05-27T01:04:55.851957+00:00"},{"alias_kind":"pith_short_12","alias_value":"OY645JZ7E3NM","created_at":"2026-05-27T01:04:55.851957+00:00"},{"alias_kind":"pith_short_16","alias_value":"OY645JZ7E3NMB4NK","created_at":"2026-05-27T01:04:55.851957+00:00"},{"alias_kind":"pith_short_8","alias_value":"OY645JZ7","created_at":"2026-05-27T01:04:55.851957+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY","json":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY.json","graph_json":"https://pith.science/api/pith-number/OY645JZ7E3NMB4NKRN3NKSCNXY/graph.json","events_json":"https://pith.science/api/pith-number/OY645JZ7E3NMB4NKRN3NKSCNXY/events.json","paper":"https://pith.science/paper/OY645JZ7"},"agent_actions":{"view_html":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY","download_json":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY.json","view_paper":"https://pith.science/paper/OY645JZ7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.02518&json=true","fetch_graph":"https://pith.science/api/pith-number/OY645JZ7E3NMB4NKRN3NKSCNXY/graph.json","fetch_events":"https://pith.science/api/pith-number/OY645JZ7E3NMB4NKRN3NKSCNXY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY/action/storage_attestation","attest_author":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY/action/author_attestation","sign_citation":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY/action/citation_signature","submit_replication":"https://pith.science/pith/OY645JZ7E3NMB4NKRN3NKSCNXY/action/replication_record"}},"created_at":"2026-05-27T01:04:55.851957+00:00","updated_at":"2026-05-27T01:04:55.851957+00:00"}