{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:42BTHFBO6733UOG5GGERTRYEDT","short_pith_number":"pith:42BTHFBO","schema_version":"1.0","canonical_sha256":"e68333942ef7f7ba38dd318919c7041cc5300b9c6c5189388e97c7cf73a1d617","source":{"kind":"arxiv","id":"2603.29875","version":3},"attestation_state":"computed","paper":{"title":"UnWeaving the knots of GraphRAG -- turns out VectorRAG is almost enough","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Entity decomposition of documents into cross-chunk links simplifies GraphRAG to near VectorRAG performance while preserving source fidelity.","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.IR","authors_text":"Adam Kozakiewicz, Mateusz Czy\\.znikiewicz, Mateusz Gali\\'nski, Micha{\\l} Godziszewski, Micha{\\l} Karpowicz, Ryszard Tuora, Tomasz Zi\\k{e}tkiewicz","submitted_at":"2026-02-06T11:37:10Z","abstract_excerpt":"One of the key problems in Retrieval-augmented generation (RAG) systems is that chunk-based retrieval pipelines represent the source chunks as atomic objects, mixing the information contained within such a chunk into a single vector. These vector representations are then fundamentally treated as isolated, independent and self-sufficient, with no attempt to represent possible relations between them. Such an approach has no dedicated mechanisms for handling multi-hop questions. Graph-based RAG systems aimed to ameliorate this problem by modeling information as knowledge-graphs, with entities rep"},"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":"2603.29875","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-02-06T11:37:10Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"ac3dff5b864ed7943b3b8ce6fc1fde67b916e92266c42eda96eea0b7fa6115d4","abstract_canon_sha256":"2c37fcac6caa56b6177af9c68f1d365c245853792da85a68903d754a17a7dec1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T02:07:25.723582Z","signature_b64":"JKGv0jxsMY6i4uVe8hUpZIBuDr5xKojGDBvBJk/E5ZabmI7mumRylJiVCG57bTa/uIs+jpz80hRafo57HtPpAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e68333942ef7f7ba38dd318919c7041cc5300b9c6c5189388e97c7cf73a1d617","last_reissued_at":"2026-06-09T02:07:25.722438Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T02:07:25.722438Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"UnWeaving the knots of GraphRAG -- turns out VectorRAG is almost enough","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Entity decomposition of documents into cross-chunk links simplifies GraphRAG to near VectorRAG performance while preserving source fidelity.","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.IR","authors_text":"Adam Kozakiewicz, Mateusz Czy\\.znikiewicz, Mateusz Gali\\'nski, Micha{\\l} Godziszewski, Micha{\\l} Karpowicz, Ryszard Tuora, Tomasz Zi\\k{e}tkiewicz","submitted_at":"2026-02-06T11:37:10Z","abstract_excerpt":"One of the key problems in Retrieval-augmented generation (RAG) systems is that chunk-based retrieval pipelines represent the source chunks as atomic objects, mixing the information contained within such a chunk into a single vector. These vector representations are then fundamentally treated as isolated, independent and self-sufficient, with no attempt to represent possible relations between them. Such an approach has no dedicated mechanisms for handling multi-hop questions. Graph-based RAG systems aimed to ameliorate this problem by modeling information as knowledge-graphs, with entities rep"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"entity-based decomposition yields a more distilled representation of original information, and additionally serves to reduce noise in the indexing, and generation process.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That an LLM can reliably extract entities across chunks without introducing systematic errors or hallucinations that would then propagate into retrieval.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"UnWeaver disentangles documents into entities via LLM to retrieve original chunks, yielding a simpler alternative to GraphRAG that still reduces noise and preserves source fidelity.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Entity decomposition of documents into cross-chunk links simplifies GraphRAG to near VectorRAG performance while preserving source fidelity.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"e69e2cb58a4dfb425287868efb97ee519c9ca3a33afcde481042e17660201771"},"source":{"id":"2603.29875","kind":"arxiv","version":3},"verdict":{"id":"80cf293e-fb16-4d78-9158-a82097b3509b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T06:55:43.692655Z","strongest_claim":"entity-based decomposition yields a more distilled representation of original information, and additionally serves to reduce noise in the indexing, and generation process.","one_line_summary":"UnWeaver disentangles documents into entities via LLM to retrieve original chunks, yielding a simpler alternative to GraphRAG that still reduces noise and preserves source fidelity.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That an LLM can reliably extract entities across chunks without introducing systematic errors or hallucinations that would then propagate into retrieval.","pith_extraction_headline":"Entity decomposition of documents into cross-chunk links simplifies GraphRAG to near VectorRAG performance while preserving source fidelity."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2603.29875/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":"2603.29875","created_at":"2026-06-09T02:07:25.722600+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.29875v3","created_at":"2026-06-09T02:07:25.722600+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.29875","created_at":"2026-06-09T02:07:25.722600+00:00"},{"alias_kind":"pith_short_12","alias_value":"42BTHFBO6733","created_at":"2026-06-09T02:07:25.722600+00:00"},{"alias_kind":"pith_short_16","alias_value":"42BTHFBO6733UOG5","created_at":"2026-06-09T02:07:25.722600+00:00"},{"alias_kind":"pith_short_8","alias_value":"42BTHFBO","created_at":"2026-06-09T02:07:25.722600+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.27164","citing_title":"Query Symbolically or Retrieve Semantically? A Dataset and Method for Semi-Structured Question Answering","ref_index":45,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT","json":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT.json","graph_json":"https://pith.science/api/pith-number/42BTHFBO6733UOG5GGERTRYEDT/graph.json","events_json":"https://pith.science/api/pith-number/42BTHFBO6733UOG5GGERTRYEDT/events.json","paper":"https://pith.science/paper/42BTHFBO"},"agent_actions":{"view_html":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT","download_json":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT.json","view_paper":"https://pith.science/paper/42BTHFBO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.29875&json=true","fetch_graph":"https://pith.science/api/pith-number/42BTHFBO6733UOG5GGERTRYEDT/graph.json","fetch_events":"https://pith.science/api/pith-number/42BTHFBO6733UOG5GGERTRYEDT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT/action/storage_attestation","attest_author":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT/action/author_attestation","sign_citation":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT/action/citation_signature","submit_replication":"https://pith.science/pith/42BTHFBO6733UOG5GGERTRYEDT/action/replication_record"}},"created_at":"2026-06-09T02:07:25.722600+00:00","updated_at":"2026-06-09T02:07:25.722600+00:00"}