{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:75LO3J3KDQSR6FC45AYFQZUXFQ","short_pith_number":"pith:75LO3J3K","schema_version":"1.0","canonical_sha256":"ff56eda76a1c251f145ce8305866972c2947cc381e224b7b33d8e028e633977c","source":{"kind":"arxiv","id":"2605.21543","version":1},"attestation_state":"computed","paper":{"title":"Provable Joint Decontamination for Benchmarking Multiple Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hao Zeng, Hongxin Wei, Zhenlong Liu","submitted_at":"2026-05-20T09:16:39Z","abstract_excerpt":"Benchmark data contamination has become a central challenge in LLM evaluation: when evaluation examples appear in the training data of one or more audited models, reported performance can be inflated and cross-model comparisons become unreliable. A broad line of training-data detection work designs scores to quantify how strongly a model memorizes a given data point, but these score-based methods lack theoretical guarantees. Recent conformal approaches provide provable false-identification control for a single model; however, applying them separately to each model can produce model-specific be"},"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":"2605.21543","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-20T09:16:39Z","cross_cats_sorted":[],"title_canon_sha256":"3bee58ea9ac08bd9e46a82a5e55e11a710250e3022159903cb42888473960cfc","abstract_canon_sha256":"69996d22e6c0f2b195bc59b9399557bbf77aa411debc6da83043f639d028acb2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T00:02:28.054515Z","signature_b64":"pHkN+6E1bTZc9SJi+GbWCx5Ps3tCSKenCY20LcWSQ9o/PWgcOVHJMeHCm2aMwEjfpQz6OtLzXFZ77t7dVAnFAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff56eda76a1c251f145ce8305866972c2947cc381e224b7b33d8e028e633977c","last_reissued_at":"2026-05-22T00:02:28.054051Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T00:02:28.054051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Provable Joint Decontamination for Benchmarking Multiple Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hao Zeng, Hongxin Wei, Zhenlong Liu","submitted_at":"2026-05-20T09:16:39Z","abstract_excerpt":"Benchmark data contamination has become a central challenge in LLM evaluation: when evaluation examples appear in the training data of one or more audited models, reported performance can be inflated and cross-model comparisons become unreliable. A broad line of training-data detection work designs scores to quantify how strongly a model memorizes a given data point, but these score-based methods lack theoretical guarantees. Recent conformal approaches provide provable false-identification control for a single model; however, applying them separately to each model can produce model-specific be"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21543","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/2605.21543/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":"2605.21543","created_at":"2026-05-22T00:02:28.054116+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.21543v1","created_at":"2026-05-22T00:02:28.054116+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21543","created_at":"2026-05-22T00:02:28.054116+00:00"},{"alias_kind":"pith_short_12","alias_value":"75LO3J3KDQSR","created_at":"2026-05-22T00:02:28.054116+00:00"},{"alias_kind":"pith_short_16","alias_value":"75LO3J3KDQSR6FC4","created_at":"2026-05-22T00:02:28.054116+00:00"},{"alias_kind":"pith_short_8","alias_value":"75LO3J3K","created_at":"2026-05-22T00:02:28.054116+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/75LO3J3KDQSR6FC45AYFQZUXFQ","json":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ.json","graph_json":"https://pith.science/api/pith-number/75LO3J3KDQSR6FC45AYFQZUXFQ/graph.json","events_json":"https://pith.science/api/pith-number/75LO3J3KDQSR6FC45AYFQZUXFQ/events.json","paper":"https://pith.science/paper/75LO3J3K"},"agent_actions":{"view_html":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ","download_json":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ.json","view_paper":"https://pith.science/paper/75LO3J3K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.21543&json=true","fetch_graph":"https://pith.science/api/pith-number/75LO3J3KDQSR6FC45AYFQZUXFQ/graph.json","fetch_events":"https://pith.science/api/pith-number/75LO3J3KDQSR6FC45AYFQZUXFQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ/action/storage_attestation","attest_author":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ/action/author_attestation","sign_citation":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ/action/citation_signature","submit_replication":"https://pith.science/pith/75LO3J3KDQSR6FC45AYFQZUXFQ/action/replication_record"}},"created_at":"2026-05-22T00:02:28.054116+00:00","updated_at":"2026-05-22T00:02:28.054116+00:00"}