{"id":"b02eb2fabe24e706f28f07fa8f199447cd2a404842044c3363c9f8a198bdd3cc","title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","argument":"Samarth Vinayaka argues that deleting facts labeled stale can lower an LLM memory system’s downstream accuracy when resolver errors remove facts needed to answer a query. In controlled sensitivity interventions across three models, hard-drop changed from gains of +0.186 to +0.230 with perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability. Paired targeted interventions found that false-drop caused significantly greater degradation than false-keep; the abstract also cautions that retaining facts does not prevent failures when incorrect labels steer generation. On a frozen 102/102 calibration/test split, a deployment-matched gate reached 0.655 accuracy, compared with 0.641 for always hard-drop and 0.662 for test-informed oracle selection. The author says global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. The practical point is that deletion carries a risk: validate query-critical false-drop risk on matched data, or use label-free retention when that validation is unavailable. The abstract does not establish how these results transfer beyond the tested models and tasks.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the abstract, which does not identify the three models or describe the tasks and interventions in enough detail to assess how broadly the results apply. The abstract does not report the measured reliability levels or explain how the deployment-matched gate is constructed.","speaker":{"name":"Vinayaka, Samarth","roleAtSource":"Article author"},"source":{"title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.17338","publishedAt":"2026-08-30","dateBasis":"Date basis not provided."},"evidence":[{"text":"The author argues that deleting stale-labeled facts can reduce downstream accuracy if resolver errors delete facts critical to a query.","excerpt":"this hard-filtering step can reduce downstream accuracy when resolver errors delete query-critical facts","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"In controlled sensitivity interventions across three models, hard-drop shifted from gains with perfect resolution to losses at the lowest tested reliability.","excerpt":"moves hard-drop from gains of +0.186 to +0.230 at perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"The paired targeted interventions found false-drop more damaging than false-keep, while incorrect labels could also steer generation under a non-deleting baseline.","excerpt":"false-drop causes significantly greater degradation than false-keep, while a non-deleting baseline can still fail when incorrect labels steer generation","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"On the stated calibration/test split, the deployment-matched gate scored below the test-informed oracle selection and above always hard-drop.","excerpt":"a simple deployment-matched gate reaches 0.655 accuracy versus 0.641 for always hard-drop and 0.662 for test-informed oracle selection","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"The author characterizes global-accuracy crossovers as diagnostics, not deployment thresholds, and recommends validating query-critical false-drop risk on matched data or defaulting to label-free retention when validation is unavailable.","excerpt":"Global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. Deletion should instead be treated as cost-sensitive","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.17338","checkedAt":"2026-11-16T07:48:11.551Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How does query-critical false-drop risk vary across models and tasks on matched data, and how well does that validation support the proposed deletion gate?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","summary":"Samarth Vinayaka argues that deleting facts labeled stale can lower an LLM memory system’s downstream accuracy when resolver errors remove facts needed to answer a query. In controlled sensitivity interventions across three models, hard-drop changed from gains of +0.186 to +0.230 with perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability. Paired targeted interventions found that false-drop caused significantly greater degradation than false-keep; the abstract also cautions that retaining facts does not prevent failures when incorrect labels steer generation. On a frozen 102/102 calibration/test split, a deployment-matched gate reached 0.655 accuracy, compared with 0.641 for always hard-drop and 0.662 for test-informed oracle selection. The author says global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. The practical point is that deletion carries a risk: validate query-critical false-drop risk on matched data, or use label-free retention when that validation is unavailable. The abstract does not establish how these results transfer beyond the tested models and tasks.","question":"How does query-critical false-drop risk vary across models and tasks on matched data, and how well does that validation support the proposed deletion gate?","claims":[{"text":"The author argues that deleting stale-labeled facts can reduce downstream accuracy if resolver errors delete facts critical to a query.","excerpt":"this hard-filtering step can reduce downstream accuracy when resolver errors delete query-critical facts","support":"source-stated"},{"text":"In controlled sensitivity interventions across three models, hard-drop shifted from gains with perfect resolution to losses at the lowest tested reliability.","excerpt":"moves hard-drop from gains of +0.186 to +0.230 at perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability","support":"source-stated"},{"text":"The paired targeted interventions found false-drop more damaging than false-keep, while incorrect labels could also steer generation under a non-deleting baseline.","excerpt":"false-drop causes significantly greater degradation than false-keep, while a non-deleting baseline can still fail when incorrect labels steer generation","support":"source-stated"},{"text":"On the stated calibration/test split, the deployment-matched gate scored below the test-informed oracle selection and above always hard-drop.","excerpt":"a simple deployment-matched gate reaches 0.655 accuracy versus 0.641 for always hard-drop and 0.662 for test-informed oracle selection","support":"source-stated"},{"text":"The author characterizes global-accuracy crossovers as diagnostics, not deployment thresholds, and recommends validating query-critical false-drop risk on matched data or defaulting to label-free retention when validation is unavailable.","excerpt":"Global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. Deletion should instead be treated as cost-sensitive","support":"source-stated"}],"limitations":["The supplied evidence is the abstract, which does not identify the three models or describe the tasks and interventions in enough detail to assess how broadly the results apply.","The abstract does not report the measured reliability levels or explain how the deployment-matched gate is constructed."],"mediaId":"f5dc4cd120b45659518fb591bab6aa27c9c58619caff9fd1deb945dff07a13a6","correctionReason":null,"id":"b02eb2fabe24e706f28f07fa8f199447cd2a404842044c3363c9f8a198bdd3cc","publicationId":"77239063-32ea-46f6-8a61-24dcf2b390a6","version":1,"sourceUrl":"https://arxiv.org/abs/2611.17338","sourceTitle":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-08-30","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-16T05:01:12.347Z","createdAt":"2026-11-16T07:33:37.555Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"f5dc4cd120b45659518fb591bab6aa27c9c58619caff9fd1deb945dff07a13a6","kind":"article","url":"https://arxiv.org/abs/2611.17338","sourceUrl":"https://arxiv.org/abs/2611.17338","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-16T07:32:09.340Z","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/2611.17338","checkedAt":"2026-11-16T07:48:11.551Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v9","mediaId":"f5dc4cd120b45659518fb591bab6aa27c9c58619caff9fd1deb945dff07a13a6","confidence":0.61,"provider":"typesafe","model":"jev-1.13.0","inputHash":"e3dc770c24d0d4014c078d75d9815d541247b8cff3a128f641f506dc6845148f","decidedAt":"2026-11-16T07:32:10.028Z"},"provenance":{"textHash":"f2cf6b9c8b2fce0a3e80182da1a81257c2296b82143305c5e71235b9ca0fd7b5","observationId":"4d8d2b93c46a71ef9a3f38b89fcaa38827ce0fcf31b494546b107594c9e273b3","retrievedAt":"2026-11-16T07:32:09.254Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"cc7d2163-b227-4fa9-88d6-ad619bcba92a"},"reviewer":{"model":"gpt-6-luna","runId":"7ab1a8e0-f299-46c0-8019-1f65f695c911"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-16T07:33:37.555Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Vinayaka, Samarth","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Vinayaka, Samarth","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Vinayaka, Samarth","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.17338","locator":"head/meta[name=author|citation_author]"},"assertionId":"9bf0cf6523c30c06eb33748f0a523adec409de99ddded1b137c613479386f5c9:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-16T07:48:11.551Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"9bf0cf6523c30c06eb33748f0a523adec409de99ddded1b137c613479386f5c9","sourceUrl":"https://arxiv.org/abs/2611.17338","contentSha256":"04897c8d9bf5522b48977189ea8f1911439c1d181b2ebe36a06a67e9cd30e118","textHash":"f2cf6b9c8b2fce0a3e80182da1a81257c2296b82143305c5e71235b9ca0fd7b5","observedAt":"2026-11-16T07:48:11.551Z"}}],"participationCoverage":{"total":1,"returned":1,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.17338","snapshotId":"9bf0cf6523c30c06eb33748f0a523adec409de99ddded1b137c613479386f5c9","contentSha256":"04897c8d9bf5522b48977189ea8f1911439c1d181b2ebe36a06a67e9cd30e118","textHash":"f2cf6b9c8b2fce0a3e80182da1a81257c2296b82143305c5e71235b9ca0fd7b5","observedAt":"2026-11-16T07:48:11.551Z","basis":"publisher-page-metadata","sourceLinks":39,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/b02eb2fabe24e706f28f07fa8f199447cd2a404842044c3363c9f8a198bdd3cc/source"}},"canonicalUrl":"https://escapemeridian.com/readings/when-should-llm-agents-forget-memory-filtering-under-query-b02eb2fa","formats":{"html":"/readings/when-should-llm-agents-forget-memory-filtering-under-query-b02eb2fa","markdown":"/readings/when-should-llm-agents-forget-memory-filtering-under-query-b02eb2fa.md","json":"/readings/when-should-llm-agents-forget-memory-filtering-under-query-b02eb2fa.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":[],"conditions":[],"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-llm","name":"Large language model","explanation":"A language model with many learned numerical settings, used to process and generate language.","referenceUrl":"https://developers.google.com/machine-learning/glossary#large-language-model"}],"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":"b02eb2fabe24e706f28f07fa8f199447cd2a404842044c3363c9f8a198bdd3cc","revision":"77239063-32ea-46f6-8a61-24dcf2b390a6","revisionKind":"publication-id","title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","url":"https://escapemeridian.com/readings/when-should-llm-agents-forget-memory-filtering-under-query-b02eb2fa"},"source":{"type":"source","id":"https://arxiv.org/abs/2611.17338","revision":"f2cf6b9c8b2fce0a3e80182da1a81257c2296b82143305c5e71235b9ca0fd7b5","revisionKind":"sha256-source-text","title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","url":"https://arxiv.org/abs/2611.17338"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings to this reading. Shared topics are not citations. Other record types need an explicit reference adapter."},"revisions":[{"id":"b02eb2fabe24e706f28f07fa8f199447cd2a404842044c3363c9f8a198bdd3cc","title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","argument":"Samarth Vinayaka argues that deleting facts labeled stale can lower an LLM memory system’s downstream accuracy when resolver errors remove facts needed to answer a query. In controlled sensitivity interventions across three models, hard-drop changed from gains of +0.186 to +0.230 with perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability. Paired targeted interventions found that false-drop caused significantly greater degradation than false-keep; the abstract also cautions that retaining facts does not prevent failures when incorrect labels steer generation. On a frozen 102/102 calibration/test split, a deployment-matched gate reached 0.655 accuracy, compared with 0.641 for always hard-drop and 0.662 for test-informed oracle selection. The author says global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. The practical point is that deletion carries a risk: validate query-critical false-drop risk on matched data, or use label-free retention when that validation is unavailable. The abstract does not establish how these results transfer beyond the tested models and tasks.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the abstract, which does not identify the three models or describe the tasks and interventions in enough detail to assess how broadly the results apply. The abstract does not report the measured reliability levels or explain how the deployment-matched gate is constructed.","speaker":{"name":"Vinayaka, Samarth","roleAtSource":"Article author"},"source":{"title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.17338","publishedAt":"2026-08-30","dateBasis":"Date basis not provided."},"evidence":[{"text":"The author argues that deleting stale-labeled facts can reduce downstream accuracy if resolver errors delete facts critical to a query.","excerpt":"this hard-filtering step can reduce downstream accuracy when resolver errors delete query-critical facts","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"In controlled sensitivity interventions across three models, hard-drop shifted from gains with perfect resolution to losses at the lowest tested reliability.","excerpt":"moves hard-drop from gains of +0.186 to +0.230 at perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"The paired targeted interventions found false-drop more damaging than false-keep, while incorrect labels could also steer generation under a non-deleting baseline.","excerpt":"false-drop causes significantly greater degradation than false-keep, while a non-deleting baseline can still fail when incorrect labels steer generation","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"On the stated calibration/test split, the deployment-matched gate scored below the test-informed oracle selection and above always hard-drop.","excerpt":"a simple deployment-matched gate reaches 0.655 accuracy versus 0.641 for always hard-drop and 0.662 for test-informed oracle selection","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"},{"text":"The author characterizes global-accuracy crossovers as diagnostics, not deployment thresholds, and recommends validating query-critical false-drop risk on matched data or defaulting to label-free retention when validation is unavailable.","excerpt":"Global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. Deletion should instead be treated as cost-sensitive","url":"https://arxiv.org/abs/2611.17338","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.17338","checkedAt":"2026-11-16T07:48:11.551Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How does query-critical false-drop risk vary across models and tasks on matched data, and how well does that validation support the proposed deletion gate?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","summary":"Samarth Vinayaka argues that deleting facts labeled stale can lower an LLM memory system’s downstream accuracy when resolver errors remove facts needed to answer a query. In controlled sensitivity interventions across three models, hard-drop changed from gains of +0.186 to +0.230 with perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability. Paired targeted interventions found that false-drop caused significantly greater degradation than false-keep; the abstract also cautions that retaining facts does not prevent failures when incorrect labels steer generation. On a frozen 102/102 calibration/test split, a deployment-matched gate reached 0.655 accuracy, compared with 0.641 for always hard-drop and 0.662 for test-informed oracle selection. The author says global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. The practical point is that deletion carries a risk: validate query-critical false-drop risk on matched data, or use label-free retention when that validation is unavailable. The abstract does not establish how these results transfer beyond the tested models and tasks.","question":"How does query-critical false-drop risk vary across models and tasks on matched data, and how well does that validation support the proposed deletion gate?","claims":[{"text":"The author argues that deleting stale-labeled facts can reduce downstream accuracy if resolver errors delete facts critical to a query.","excerpt":"this hard-filtering step can reduce downstream accuracy when resolver errors delete query-critical facts","support":"source-stated"},{"text":"In controlled sensitivity interventions across three models, hard-drop shifted from gains with perfect resolution to losses at the lowest tested reliability.","excerpt":"moves hard-drop from gains of +0.186 to +0.230 at perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability","support":"source-stated"},{"text":"The paired targeted interventions found false-drop more damaging than false-keep, while incorrect labels could also steer generation under a non-deleting baseline.","excerpt":"false-drop causes significantly greater degradation than false-keep, while a non-deleting baseline can still fail when incorrect labels steer generation","support":"source-stated"},{"text":"On the stated calibration/test split, the deployment-matched gate scored below the test-informed oracle selection and above always hard-drop.","excerpt":"a simple deployment-matched gate reaches 0.655 accuracy versus 0.641 for always hard-drop and 0.662 for test-informed oracle selection","support":"source-stated"},{"text":"The author characterizes global-accuracy crossovers as diagnostics, not deployment thresholds, and recommends validating query-critical false-drop risk on matched data or defaulting to label-free retention when validation is unavailable.","excerpt":"Global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. Deletion should instead be treated as cost-sensitive","support":"source-stated"}],"limitations":["The supplied evidence is the abstract, which does not identify the three models or describe the tasks and interventions in enough detail to assess how broadly the results apply.","The abstract does not report the measured reliability levels or explain how the deployment-matched gate is constructed."],"mediaId":"f5dc4cd120b45659518fb591bab6aa27c9c58619caff9fd1deb945dff07a13a6","correctionReason":null,"id":"b02eb2fabe24e706f28f07fa8f199447cd2a404842044c3363c9f8a198bdd3cc","publicationId":"77239063-32ea-46f6-8a61-24dcf2b390a6","version":1,"sourceUrl":"https://arxiv.org/abs/2611.17338","sourceTitle":"When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-08-30","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-16T05:01:12.347Z","createdAt":"2026-11-16T07:33:37.555Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"f5dc4cd120b45659518fb591bab6aa27c9c58619caff9fd1deb945dff07a13a6","kind":"article","url":"https://arxiv.org/abs/2611.17338","sourceUrl":"https://arxiv.org/abs/2611.17338","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-16T07:32:09.340Z","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/2611.17338","checkedAt":"2026-11-16T07:48:11.551Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v9","mediaId":"f5dc4cd120b45659518fb591bab6aa27c9c58619caff9fd1deb945dff07a13a6","confidence":0.61,"provider":"typesafe","model":"jev-1.13.0","inputHash":"e3dc770c24d0d4014c078d75d9815d541247b8cff3a128f641f506dc6845148f","decidedAt":"2026-11-16T07:32:10.028Z"},"provenance":{"textHash":"f2cf6b9c8b2fce0a3e80182da1a81257c2296b82143305c5e71235b9ca0fd7b5","observationId":"4d8d2b93c46a71ef9a3f38b89fcaa38827ce0fcf31b494546b107594c9e273b3","retrievedAt":"2026-11-16T07:32:09.254Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"cc7d2163-b227-4fa9-88d6-ad619bcba92a"},"reviewer":{"model":"gpt-6-luna","runId":"7ab1a8e0-f299-46c0-8019-1f65f695c911"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-16T07:33:37.555Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Vinayaka, Samarth","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Vinayaka, Samarth","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Vinayaka, Samarth","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.17338","locator":"head/meta[name=author|citation_author]"},"assertionId":"9bf0cf6523c30c06eb33748f0a523adec409de99ddded1b137c613479386f5c9:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-16T07:48:11.551Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"9bf0cf6523c30c06eb33748f0a523adec409de99ddded1b137c613479386f5c9","sourceUrl":"https://arxiv.org/abs/2611.17338","contentSha256":"04897c8d9bf5522b48977189ea8f1911439c1d181b2ebe36a06a67e9cd30e118","textHash":"f2cf6b9c8b2fce0a3e80182da1a81257c2296b82143305c5e71235b9ca0fd7b5","observedAt":"2026-11-16T07:48:11.551Z"}}],"participationCoverage":{"total":1,"returned":1,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.17338","snapshotId":"9bf0cf6523c30c06eb33748f0a523adec409de99ddded1b137c613479386f5c9","contentSha256":"04897c8d9bf5522b48977189ea8f1911439c1d181b2ebe36a06a67e9cd30e118","textHash":"f2cf6b9c8b2fce0a3e80182da1a81257c2296b82143305c5e71235b9ca0fd7b5","observedAt":"2026-11-16T07:48:11.551Z","basis":"publisher-page-metadata","sourceLinks":39,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/b02eb2fabe24e706f28f07fa8f199447cd2a404842044c3363c9f8a198bdd3cc/source"}}}],"audits":[{"publicationId":"77239063-32ea-46f6-8a61-24dcf2b390a6","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading correctly attributes the paper and its claims to Samarth Vinayaka, the listed sole author. Its findings and recommendations are framed as the author's arguments and the abstract's results, not as established facts.","excerpt":"Authors:Samarth Vinayaka"},{"dimension":"summary","status":"uncertain","explanation":"The summary closely follows the abstract, but it leaves the central jargon unexplained for a general reader: resolver errors, hard-drop, false-drop vs false-keep, query-critical, deployment-matched gate, test-informed oracle selection, gold labels, global-accuracy crossovers, label-free retention and the 102/102 calibration/test split. It never says in plain terms that a mistaken stale label can delete a fact still needed to answer a question. Without that, the significance of the numbers is hard to grasp. Smallest repair: briefly explain these terms, or add term-library tooltips, especially resolver, false-drop/false-keep and the gate.","excerpt":"LLM memory systems often delete facts classified as stale."},{"dimension":"date","status":"pass","explanation":"originalPublished is 2026-08-30, which matches the arXiv v1 submission date. The 2026-11-16 feed date was correctly not used.","excerpt":"[Submitted on 30 Aug 2026]"},{"dimension":"support","status":"pass","explanation":"Every numeric claim and comparison matches the abstract: the gains and losses, false-drop vs false-keep, the non-deleting baseline failure, and the 0.655, 0.641 and 0.662 accuracies. The crossover caveat and the recommendation are also accurate. The draft does not overstate the gate's small advantage. Its stated limitations and question follow from the evidence. A minor wording point: the source names a single controlled sensitivity intervention, while the draft uses the plural; this is not material.","excerpt":"validate query-critical false-drop risk on matched data, or default to label-free retention when such validation is unavailable"},{"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":"2026-11-16T07:33:37.555Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}