{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZM3OD6UNANXV63OP35FJTJHVJH","short_pith_number":"pith:ZM3OD6UN","schema_version":"1.0","canonical_sha256":"cb36e1fa8d036f5f6dcfdf4a99a4f549d8ce164cfbccbbed5d217c7103c20726","source":{"kind":"arxiv","id":"2606.03305","version":1},"attestation_state":"computed","paper":{"title":"The Reliability Gap in Benchmark Auditing: Distribution Shift and Scale as Failure Modes of Contamination Detection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jan Dubi\\'nski, Sebastian Cygert, Wojciech Zarzecki","submitted_at":"2026-06-02T08:21:22Z","abstract_excerpt":"Benchmark contamination, where evaluation examples appear in a model's training data, threatens the validity of LLM assessment. Statistical tools for detecting training-data membership exist, but have been validated almost exclusively in controlled academic regimes: large, homogeneous pre-training corpora and transparent, single-stage training pipelines. Whether these methods remain reliable in realistic auditing scenarios remains unclear. We identify two under-studied failure modes: distribution shift, which arises when suspect and validation sets violate the IID assumption, and scale constra"},"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":"2606.03305","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-02T08:21:22Z","cross_cats_sorted":[],"title_canon_sha256":"3b442cade36879c90c8a31349fcc23a9488bfb27a577c7cafe7d87d8f3f57631","abstract_canon_sha256":"a2826931c24433dca81b023f1e6498037a9fd4178ce5da67fe7c33483170f200"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T01:05:54.537290Z","signature_b64":"PSLsimvXOELpQAWYrV3DY4Agzon/LnBSzP1xP4cDl58FjYu46O++EaMC8xSV2qO+bbeFUmvpFseGIqAgqyrzCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cb36e1fa8d036f5f6dcfdf4a99a4f549d8ce164cfbccbbed5d217c7103c20726","last_reissued_at":"2026-06-03T01:05:54.536895Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T01:05:54.536895Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Reliability Gap in Benchmark Auditing: Distribution Shift and Scale as Failure Modes of Contamination Detection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jan Dubi\\'nski, Sebastian Cygert, Wojciech Zarzecki","submitted_at":"2026-06-02T08:21:22Z","abstract_excerpt":"Benchmark contamination, where evaluation examples appear in a model's training data, threatens the validity of LLM assessment. Statistical tools for detecting training-data membership exist, but have been validated almost exclusively in controlled academic regimes: large, homogeneous pre-training corpora and transparent, single-stage training pipelines. Whether these methods remain reliable in realistic auditing scenarios remains unclear. We identify two under-studied failure modes: distribution shift, which arises when suspect and validation sets violate the IID assumption, and scale constra"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.03305","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/2606.03305/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":"2606.03305","created_at":"2026-06-03T01:05:54.536956+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.03305v1","created_at":"2026-06-03T01:05:54.536956+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.03305","created_at":"2026-06-03T01:05:54.536956+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZM3OD6UNANXV","created_at":"2026-06-03T01:05:54.536956+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZM3OD6UNANXV63OP","created_at":"2026-06-03T01:05:54.536956+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZM3OD6UN","created_at":"2026-06-03T01:05:54.536956+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/ZM3OD6UNANXV63OP35FJTJHVJH","json":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH.json","graph_json":"https://pith.science/api/pith-number/ZM3OD6UNANXV63OP35FJTJHVJH/graph.json","events_json":"https://pith.science/api/pith-number/ZM3OD6UNANXV63OP35FJTJHVJH/events.json","paper":"https://pith.science/paper/ZM3OD6UN"},"agent_actions":{"view_html":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH","download_json":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH.json","view_paper":"https://pith.science/paper/ZM3OD6UN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.03305&json=true","fetch_graph":"https://pith.science/api/pith-number/ZM3OD6UNANXV63OP35FJTJHVJH/graph.json","fetch_events":"https://pith.science/api/pith-number/ZM3OD6UNANXV63OP35FJTJHVJH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH/action/storage_attestation","attest_author":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH/action/author_attestation","sign_citation":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH/action/citation_signature","submit_replication":"https://pith.science/pith/ZM3OD6UNANXV63OP35FJTJHVJH/action/replication_record"}},"created_at":"2026-06-03T01:05:54.536956+00:00","updated_at":"2026-06-03T01:05:54.536956+00:00"}