{"id":"8b88f1920487eac3a062036da58e73f6631abc819802a24c38fdd21645e6073b","title":"arXiv cs.AI: Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","argument":"Uday Allu presents Lightfoot, a GraphRAG system designed to make building and deleting documents from a graph-based retrieval system less costly. Allu says each entity, link, fact and synonym edge comes deterministically from passage-level extractions with chunk-level provenance. Indexing uses one language-model call per passage, retrieval uses none, and deleting a document is a database operation with no model calls. On the 11,402-passage MuSiQue corpus, the paper reports that deleting 10% of documents left a graph identical to a fresh index of the surviving documents. In comparisons using the same extraction, answer and embedding models, Lightfoot built the cheapest index across MuSiQue, 2WikiMultiHopQA and HotpotQA. Its results against HippoRAG 2 varied by task: Lightfoot did better on MuSiQue retrieval and HotpotQA answers, but worse on 2WikiMultiHopQA. The trade-off matters because lower indexing costs and straightforward deletion do not mean better retrieval or answers on every dataset. These results are from the named benchmarks; the abstract does not establish how Lightfoot performs in deployed systems.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The evidence is the arXiv abstract; it summarizes benchmark results but does not provide the paper’s methods or detailed evaluation context. The reported results concern MuSiQue, 2WikiMultiHopQA and HotpotQA; the abstract does not establish performance in deployed systems or on other datasets.","speaker":{"name":"Allu, Uday","roleAtSource":"Article author"},"source":{"title":"Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.11010","publishedAt":"2026-11-07","dateBasis":"Date basis not provided."},"evidence":[{"text":"Uday Allu presents Lightfoot as a GraphRAG system whose graph elements are deterministic functions of passage extractions carrying chunk-level provenance.","excerpt":"every entity, link, fact and synonym edge is a deterministic function of per-passage extractions carrying chunk-level provenance","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"The paper says indexing requires one language-model call per passage, retrieval requires none, and document deletion requires no model calls.","excerpt":"Indexing makes one language-model call per passage, retrieval makes none, and deleting a document is a database operation with no model calls.","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"On MuSiQue, the authors report that after deleting 10% of documents the resulting graph matched a fresh index of the surviving documents.","excerpt":"after deleting 10% of the documents, the graph is identical to a fresh index of the surviving documents.","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"Lightfoot had the cheapest index across the three named datasets, but its comparison with HippoRAG 2 varied by dataset and task.","excerpt":"Lightfoot builds the cheapest index on all three datasets, 26-31% below HippoRAG 2","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"The abstract reports better Lightfoot results than HippoRAG 2 on MuSiQue retrieval and HotpotQA answers, and worse results on 2WikiMultiHopQA.","excerpt":"significantly better on MuSiQue retrieval (+2.8 points recall@5 and +4.3 recall@10 on a held-out split) and on HotpotQA answers (+7.7 points exact match), and significantly worse on 2WikiMultiHopQA (-3.9 recall@2, -8.4 F1).","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.11010","checkedAt":"2026-11-12T08:06:08.342Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Would Lightfoot’s reported deletion behavior and benchmark results hold on other datasets or in deployed systems?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","summary":"Uday Allu presents Lightfoot, a GraphRAG system designed to make building and deleting documents from a graph-based retrieval system less costly. Allu says each entity, link, fact and synonym edge comes deterministically from passage-level extractions with chunk-level provenance. Indexing uses one language-model call per passage, retrieval uses none, and deleting a document is a database operation with no model calls. On the 11,402-passage MuSiQue corpus, the paper reports that deleting 10% of documents left a graph identical to a fresh index of the surviving documents. In comparisons using the same extraction, answer and embedding models, Lightfoot built the cheapest index across MuSiQue, 2WikiMultiHopQA and HotpotQA. Its results against HippoRAG 2 varied by task: Lightfoot did better on MuSiQue retrieval and HotpotQA answers, but worse on 2WikiMultiHopQA. The trade-off matters because lower indexing costs and straightforward deletion do not mean better retrieval or answers on every dataset. These results are from the named benchmarks; the abstract does not establish how Lightfoot performs in deployed systems.","question":"Would Lightfoot’s reported deletion behavior and benchmark results hold on other datasets or in deployed systems?","claims":[{"text":"Uday Allu presents Lightfoot as a GraphRAG system whose graph elements are deterministic functions of passage extractions carrying chunk-level provenance.","excerpt":"every entity, link, fact and synonym edge is a deterministic function of per-passage extractions carrying chunk-level provenance","support":"source-stated"},{"text":"The paper says indexing requires one language-model call per passage, retrieval requires none, and document deletion requires no model calls.","excerpt":"Indexing makes one language-model call per passage, retrieval makes none, and deleting a document is a database operation with no model calls.","support":"source-stated"},{"text":"On MuSiQue, the authors report that after deleting 10% of documents the resulting graph matched a fresh index of the surviving documents.","excerpt":"after deleting 10% of the documents, the graph is identical to a fresh index of the surviving documents.","support":"source-stated"},{"text":"Lightfoot had the cheapest index across the three named datasets, but its comparison with HippoRAG 2 varied by dataset and task.","excerpt":"Lightfoot builds the cheapest index on all three datasets, 26-31% below HippoRAG 2","support":"source-stated"},{"text":"The abstract reports better Lightfoot results than HippoRAG 2 on MuSiQue retrieval and HotpotQA answers, and worse results on 2WikiMultiHopQA.","excerpt":"significantly better on MuSiQue retrieval (+2.8 points recall@5 and +4.3 recall@10 on a held-out split) and on HotpotQA answers (+7.7 points exact match), and significantly worse on 2WikiMultiHopQA (-3.9 recall@2, -8.4 F1).","support":"source-stated"}],"limitations":["The evidence is the arXiv abstract; it summarizes benchmark results but does not provide the paper’s methods or detailed evaluation context.","The reported results concern MuSiQue, 2WikiMultiHopQA and HotpotQA; the abstract does not establish performance in deployed systems or on other datasets."],"mediaId":"8a2ec17ce2845743dd095d691efd2207151b1d9e52bcefe80d6562c6f0d407a7","correctionReason":null,"id":"8b88f1920487eac3a062036da58e73f6631abc819802a24c38fdd21645e6073b","publicationId":"7be5e36e-6012-4034-8f55-e6b190e31304","version":1,"sourceUrl":"https://arxiv.org/abs/2611.11010","sourceTitle":"Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-07","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-10T08:36:39.110Z","createdAt":"2026-11-12T07:02:42.915Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"8a2ec17ce2845743dd095d691efd2207151b1d9e52bcefe80d6562c6f0d407a7","kind":"article","url":"https://arxiv.org/abs/2611.11010","sourceUrl":"https://arxiv.org/abs/2611.11010","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-12T07:01:53.705Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.11010","checkedAt":"2026-11-12T08:06:08.342Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v6","mediaId":"8a2ec17ce2845743dd095d691efd2207151b1d9e52bcefe80d6562c6f0d407a7","confidence":0.65,"provider":"typesafe","model":"jev-1.13.0","inputHash":"fec027b908832436603409dd3b62d78645096b49145963047758586a5970b697","decidedAt":"2026-11-12T07:01:54.432Z"},"provenance":{"textHash":"aae817ca98d66898565e745352ad3abddc9ae999073d7d74ccb17f70ce6f07bc","observationId":"4ed9d420480ccd886703c58ca0c49b0b02985c21c297cbd04bf7694ac2bbb341","retrievedAt":"2026-11-12T07:01:53.627Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"52a174a4-ddd6-4aa5-8b3b-5540c33f7b5b"},"reviewer":{"model":"gpt-6-luna","runId":"7e487bcd-0274-4ead-82aa-01d63815552d"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-12T07:02:42.915Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Allu, Uday","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Allu, Uday","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Allu, Uday","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.11010","locator":"head/meta[name=author|citation_author]"},"assertionId":"37aa61a34a9bbd1c0171a3fdfc333159bf91d56442ff55a30dfc60d3fe1a1fd3:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T08:06:08.342Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"37aa61a34a9bbd1c0171a3fdfc333159bf91d56442ff55a30dfc60d3fe1a1fd3","sourceUrl":"https://arxiv.org/abs/2611.11010","contentSha256":"e20087a56fea2f4c6f58695f19808496fff09595f10e1789d225b5278a5b1525","textHash":"aae817ca98d66898565e745352ad3abddc9ae999073d7d74ccb17f70ce6f07bc","observedAt":"2026-11-12T08:06:08.342Z"}}],"participationCoverage":{"total":1,"returned":1,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.11010","snapshotId":"37aa61a34a9bbd1c0171a3fdfc333159bf91d56442ff55a30dfc60d3fe1a1fd3","contentSha256":"e20087a56fea2f4c6f58695f19808496fff09595f10e1789d225b5278a5b1525","textHash":"aae817ca98d66898565e745352ad3abddc9ae999073d7d74ccb17f70ce6f07bc","observedAt":"2026-11-12T08:06:08.342Z","basis":"publisher-page-metadata","sourceLinks":42,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/8b88f1920487eac3a062036da58e73f6631abc819802a24c38fdd21645e6073b/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-cs-ai-lightfoot-graph-retrieval-augmented-generation-with-8b88f192","formats":{"html":"/readings/arxiv-cs-ai-lightfoot-graph-retrieval-augmented-generation-with-8b88f192","markdown":"/readings/arxiv-cs-ai-lightfoot-graph-retrieval-augmented-generation-with-8b88f192.md","json":"/readings/arxiv-cs-ai-lightfoot-graph-retrieval-augmented-generation-with-8b88f192.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-rag","name":"Retrieval-augmented generation","explanation":"A setup in which a model first looks up relevant documents and then answers using them.","referenceUrl":"https://developers.google.com/machine-learning/glossary#retrieval-augmented-generation-rag"},{"id":"term-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"8b88f1920487eac3a062036da58e73f6631abc819802a24c38fdd21645e6073b","revision":"7be5e36e-6012-4034-8f55-e6b190e31304","revisionKind":"publication-id","title":"arXiv cs.AI: Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","url":"https://escapemeridian.com/readings/arxiv-cs-ai-lightfoot-graph-retrieval-augmented-generation-with-8b88f192"},"source":{"type":"source","id":"https://arxiv.org/abs/2611.11010","revision":"aae817ca98d66898565e745352ad3abddc9ae999073d7d74ccb17f70ce6f07bc","revisionKind":"sha256-source-text","title":"Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","url":"https://arxiv.org/abs/2611.11010"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"8b88f1920487eac3a062036da58e73f6631abc819802a24c38fdd21645e6073b","title":"arXiv cs.AI: Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","argument":"Uday Allu presents Lightfoot, a GraphRAG system designed to make building and deleting documents from a graph-based retrieval system less costly. Allu says each entity, link, fact and synonym edge comes deterministically from passage-level extractions with chunk-level provenance. Indexing uses one language-model call per passage, retrieval uses none, and deleting a document is a database operation with no model calls. On the 11,402-passage MuSiQue corpus, the paper reports that deleting 10% of documents left a graph identical to a fresh index of the surviving documents. In comparisons using the same extraction, answer and embedding models, Lightfoot built the cheapest index across MuSiQue, 2WikiMultiHopQA and HotpotQA. Its results against HippoRAG 2 varied by task: Lightfoot did better on MuSiQue retrieval and HotpotQA answers, but worse on 2WikiMultiHopQA. The trade-off matters because lower indexing costs and straightforward deletion do not mean better retrieval or answers on every dataset. These results are from the named benchmarks; the abstract does not establish how Lightfoot performs in deployed systems.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The evidence is the arXiv abstract; it summarizes benchmark results but does not provide the paper’s methods or detailed evaluation context. The reported results concern MuSiQue, 2WikiMultiHopQA and HotpotQA; the abstract does not establish performance in deployed systems or on other datasets.","speaker":{"name":"Allu, Uday","roleAtSource":"Article author"},"source":{"title":"Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.11010","publishedAt":"2026-11-07","dateBasis":"Date basis not provided."},"evidence":[{"text":"Uday Allu presents Lightfoot as a GraphRAG system whose graph elements are deterministic functions of passage extractions carrying chunk-level provenance.","excerpt":"every entity, link, fact and synonym edge is a deterministic function of per-passage extractions carrying chunk-level provenance","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"The paper says indexing requires one language-model call per passage, retrieval requires none, and document deletion requires no model calls.","excerpt":"Indexing makes one language-model call per passage, retrieval makes none, and deleting a document is a database operation with no model calls.","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"On MuSiQue, the authors report that after deleting 10% of documents the resulting graph matched a fresh index of the surviving documents.","excerpt":"after deleting 10% of the documents, the graph is identical to a fresh index of the surviving documents.","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"Lightfoot had the cheapest index across the three named datasets, but its comparison with HippoRAG 2 varied by dataset and task.","excerpt":"Lightfoot builds the cheapest index on all three datasets, 26-31% below HippoRAG 2","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"},{"text":"The abstract reports better Lightfoot results than HippoRAG 2 on MuSiQue retrieval and HotpotQA answers, and worse results on 2WikiMultiHopQA.","excerpt":"significantly better on MuSiQue retrieval (+2.8 points recall@5 and +4.3 recall@10 on a held-out split) and on HotpotQA answers (+7.7 points exact match), and significantly worse on 2WikiMultiHopQA (-3.9 recall@2, -8.4 F1).","url":"https://arxiv.org/abs/2611.11010","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.11010","checkedAt":"2026-11-12T08:06:08.342Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Would Lightfoot’s reported deletion behavior and benchmark results hold on other datasets or in deployed systems?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","summary":"Uday Allu presents Lightfoot, a GraphRAG system designed to make building and deleting documents from a graph-based retrieval system less costly. Allu says each entity, link, fact and synonym edge comes deterministically from passage-level extractions with chunk-level provenance. Indexing uses one language-model call per passage, retrieval uses none, and deleting a document is a database operation with no model calls. On the 11,402-passage MuSiQue corpus, the paper reports that deleting 10% of documents left a graph identical to a fresh index of the surviving documents. In comparisons using the same extraction, answer and embedding models, Lightfoot built the cheapest index across MuSiQue, 2WikiMultiHopQA and HotpotQA. Its results against HippoRAG 2 varied by task: Lightfoot did better on MuSiQue retrieval and HotpotQA answers, but worse on 2WikiMultiHopQA. The trade-off matters because lower indexing costs and straightforward deletion do not mean better retrieval or answers on every dataset. These results are from the named benchmarks; the abstract does not establish how Lightfoot performs in deployed systems.","question":"Would Lightfoot’s reported deletion behavior and benchmark results hold on other datasets or in deployed systems?","claims":[{"text":"Uday Allu presents Lightfoot as a GraphRAG system whose graph elements are deterministic functions of passage extractions carrying chunk-level provenance.","excerpt":"every entity, link, fact and synonym edge is a deterministic function of per-passage extractions carrying chunk-level provenance","support":"source-stated"},{"text":"The paper says indexing requires one language-model call per passage, retrieval requires none, and document deletion requires no model calls.","excerpt":"Indexing makes one language-model call per passage, retrieval makes none, and deleting a document is a database operation with no model calls.","support":"source-stated"},{"text":"On MuSiQue, the authors report that after deleting 10% of documents the resulting graph matched a fresh index of the surviving documents.","excerpt":"after deleting 10% of the documents, the graph is identical to a fresh index of the surviving documents.","support":"source-stated"},{"text":"Lightfoot had the cheapest index across the three named datasets, but its comparison with HippoRAG 2 varied by dataset and task.","excerpt":"Lightfoot builds the cheapest index on all three datasets, 26-31% below HippoRAG 2","support":"source-stated"},{"text":"The abstract reports better Lightfoot results than HippoRAG 2 on MuSiQue retrieval and HotpotQA answers, and worse results on 2WikiMultiHopQA.","excerpt":"significantly better on MuSiQue retrieval (+2.8 points recall@5 and +4.3 recall@10 on a held-out split) and on HotpotQA answers (+7.7 points exact match), and significantly worse on 2WikiMultiHopQA (-3.9 recall@2, -8.4 F1).","support":"source-stated"}],"limitations":["The evidence is the arXiv abstract; it summarizes benchmark results but does not provide the paper’s methods or detailed evaluation context.","The reported results concern MuSiQue, 2WikiMultiHopQA and HotpotQA; the abstract does not establish performance in deployed systems or on other datasets."],"mediaId":"8a2ec17ce2845743dd095d691efd2207151b1d9e52bcefe80d6562c6f0d407a7","correctionReason":null,"id":"8b88f1920487eac3a062036da58e73f6631abc819802a24c38fdd21645e6073b","publicationId":"7be5e36e-6012-4034-8f55-e6b190e31304","version":1,"sourceUrl":"https://arxiv.org/abs/2611.11010","sourceTitle":"Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-07","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-10T08:36:39.110Z","createdAt":"2026-11-12T07:02:42.915Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"8a2ec17ce2845743dd095d691efd2207151b1d9e52bcefe80d6562c6f0d407a7","kind":"article","url":"https://arxiv.org/abs/2611.11010","sourceUrl":"https://arxiv.org/abs/2611.11010","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-12T07:01:53.705Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.11010","checkedAt":"2026-11-12T08:06:08.342Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v6","mediaId":"8a2ec17ce2845743dd095d691efd2207151b1d9e52bcefe80d6562c6f0d407a7","confidence":0.65,"provider":"typesafe","model":"jev-1.13.0","inputHash":"fec027b908832436603409dd3b62d78645096b49145963047758586a5970b697","decidedAt":"2026-11-12T07:01:54.432Z"},"provenance":{"textHash":"aae817ca98d66898565e745352ad3abddc9ae999073d7d74ccb17f70ce6f07bc","observationId":"4ed9d420480ccd886703c58ca0c49b0b02985c21c297cbd04bf7694ac2bbb341","retrievedAt":"2026-11-12T07:01:53.627Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"52a174a4-ddd6-4aa5-8b3b-5540c33f7b5b"},"reviewer":{"model":"gpt-6-luna","runId":"7e487bcd-0274-4ead-82aa-01d63815552d"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-12T07:02:42.915Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Allu, Uday","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Allu, Uday","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Allu, Uday","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.11010","locator":"head/meta[name=author|citation_author]"},"assertionId":"37aa61a34a9bbd1c0171a3fdfc333159bf91d56442ff55a30dfc60d3fe1a1fd3:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T08:06:08.342Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"37aa61a34a9bbd1c0171a3fdfc333159bf91d56442ff55a30dfc60d3fe1a1fd3","sourceUrl":"https://arxiv.org/abs/2611.11010","contentSha256":"e20087a56fea2f4c6f58695f19808496fff09595f10e1789d225b5278a5b1525","textHash":"aae817ca98d66898565e745352ad3abddc9ae999073d7d74ccb17f70ce6f07bc","observedAt":"2026-11-12T08:06:08.342Z"}}],"participationCoverage":{"total":1,"returned":1,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.11010","snapshotId":"37aa61a34a9bbd1c0171a3fdfc333159bf91d56442ff55a30dfc60d3fe1a1fd3","contentSha256":"e20087a56fea2f4c6f58695f19808496fff09595f10e1789d225b5278a5b1525","textHash":"aae817ca98d66898565e745352ad3abddc9ae999073d7d74ccb17f70ce6f07bc","observedAt":"2026-11-12T08:06:08.342Z","basis":"publisher-page-metadata","sourceLinks":42,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/8b88f1920487eac3a062036da58e73f6631abc819802a24c38fdd21645e6073b/source"}}}],"audits":[{"publicationId":"7be5e36e-6012-4034-8f55-e6b190e31304","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading correctly attributes Lightfoot and its claims to Uday Allu, the sole listed author, and keeps results framed as reported. One claim says 'the authors report' although there is a single author; this is a minor wording slip, not a misattribution. Smallest repair: change it to 'the paper reports'.","excerpt":"Authors:Uday Allu"},{"dimension":"summary","status":"uncertain","explanation":"Central terms are not explained in plain language: GraphRAG, multi-hop question answering, indexing and chunk-level provenance. A non-specialist may not see why deleting documents matters. The reading also drops the paper's sharpest deletion contrast: LightRAG took 2,187 model calls and 3.7 hours to delete the same documents, and fast-graphrag has no deletion path. It also omits the motivation that research comparisons rarely measure these costs. Smallest repair: add a plain gloss of GraphRAG and deletion, plus one sentence giving the LightRAG and fast-graphrag deletion comparison.","excerpt":"Deleting the same documents took LightRAG 2,187 model calls and 3.7 hours; fast-graphrag has no deletion path."},{"dimension":"date","status":"pass","explanation":"originalPublished 2026-11-07 matches the arXiv v1 submission date. It is not the later feed date.","excerpt":"[Submitted on 7 Nov 2026]"},{"dimension":"support","status":"pass","explanation":"Every claim matches the abstract, including the per-passage model calls, the identical graph after 10% deletion, the cheapest index and the mixed results against HippoRAG 2. The trade-off sentence and the deployment caveat are appropriately conditional. The question is open, not leading. No benefits or costs are invented.","excerpt":"Lightfoot builds the cheapest index on all three datasets, 26-31% below HippoRAG 2 and 3-6x below LightRAG and fast-graphrag."},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct arXiv abstract page for this paper, which is the most direct access to the source.","excerpt":null}],"createdAt":"2026-11-12T07:02:42.915Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}