{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:5FP2D5NO6JDBKAC4XRRYNIYNKW","short_pith_number":"pith:5FP2D5NO","schema_version":"1.0","canonical_sha256":"e95fa1f5aef24615005cbc6386a30d55b5aa0db6ba538e06847384deb636bd64","source":{"kind":"arxiv","id":"2311.09783","version":2},"attestation_state":"computed","paper":{"title":"Investigating Data Contamination in Modern Benchmarks for Large Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Arman Cohan, Chunyuan Deng, Mark Gerstein, Xiangru Tang, Yilun Zhao","submitted_at":"2023-11-16T11:03:04Z","abstract_excerpt":"Recent observations have underscored a disparity between the inflated benchmark scores and the actual performance of LLMs, raising concerns about potential contamination of evaluation benchmarks. This issue is especially critical for closed-source models and certain open-source models where training data transparency is lacking. In this paper we study data contamination by proposing two methods tailored for both open-source and proprietary LLMs. We first introduce a retrieval-based system to explore potential overlaps between evaluation benchmarks and pretraining corpora. We further present a "},"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":"2311.09783","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-16T11:03:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8e8be7923f90e8911e71913086f49241ca53f97d948426fa8eecb073a5cc3484","abstract_canon_sha256":"4bc2c0a4316b5214693f3e05734176b6eaaeef5cec4598507dc573f103ec9ef6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:04:16.024884Z","signature_b64":"MbQokWUl3HCJZlgFqdgWcIEaSTS+38y8yTHWODigA3XwVuVjWLMP3Hxh5l9ayfL8UX2ths+BdFZp9CVbd0AdAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e95fa1f5aef24615005cbc6386a30d55b5aa0db6ba538e06847384deb636bd64","last_reissued_at":"2026-07-05T08:04:16.024474Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:04:16.024474Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Investigating Data Contamination in Modern Benchmarks for Large Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Arman Cohan, Chunyuan Deng, Mark Gerstein, Xiangru Tang, Yilun Zhao","submitted_at":"2023-11-16T11:03:04Z","abstract_excerpt":"Recent observations have underscored a disparity between the inflated benchmark scores and the actual performance of LLMs, raising concerns about potential contamination of evaluation benchmarks. This issue is especially critical for closed-source models and certain open-source models where training data transparency is lacking. In this paper we study data contamination by proposing two methods tailored for both open-source and proprietary LLMs. We first introduce a retrieval-based system to explore potential overlaps between evaluation benchmarks and pretraining corpora. We further present a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.09783","kind":"arxiv","version":2},"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/2311.09783/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":"2311.09783","created_at":"2026-07-05T08:04:16.024535+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.09783v2","created_at":"2026-07-05T08:04:16.024535+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.09783","created_at":"2026-07-05T08:04:16.024535+00:00"},{"alias_kind":"pith_short_12","alias_value":"5FP2D5NO6JDB","created_at":"2026-07-05T08:04:16.024535+00:00"},{"alias_kind":"pith_short_16","alias_value":"5FP2D5NO6JDBKAC4","created_at":"2026-07-05T08:04:16.024535+00:00"},{"alias_kind":"pith_short_8","alias_value":"5FP2D5NO","created_at":"2026-07-05T08:04:16.024535+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":12,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03305","citing_title":"The Reliability Gap in Benchmark Auditing: Distribution Shift and Scale as Failure Modes of Contamination Detection","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18284","citing_title":"Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier","ref_index":59,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00053","citing_title":"SWE-Router: Routing in Multi-turn Agentic Software Engineering Tasks","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03305","citing_title":"The Reliability Gap in Benchmark Auditing: Distribution Shift and Scale as Failure Modes of Contamination Detection","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26133","citing_title":"Pretraining Data Exposure in Large Language Models: A Survey of Membership Inference, Data Contamination, and Security Implications","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2406.04244","citing_title":"Benchmark Data Contamination of Large Language Models: A Survey","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2504.16155","citing_title":"PRIMETIME : Limits of LLMs in Temporal Primitives","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2507.22359","citing_title":"League of LLMs: A Benchmark-Free Paradigm for Mutual Evaluation of Large Language Models","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2511.01188","citing_title":"ZoFia: Zero-Shot Fake News Detection with Entity-Guided Retrieval and Multi-LLM Interaction","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2406.19314","citing_title":"LiveBench: A Challenging, Contamination-Limited LLM Benchmark","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08522","citing_title":"Coordinates of Capability: A Unified MTMM-Geometric Framework for LLM Evaluation","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20273","citing_title":"ActuBench: A Multi-Agent LLM Pipeline for Generation and Evaluation of Actuarial Reasoning Tasks","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW","json":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW.json","graph_json":"https://pith.science/api/pith-number/5FP2D5NO6JDBKAC4XRRYNIYNKW/graph.json","events_json":"https://pith.science/api/pith-number/5FP2D5NO6JDBKAC4XRRYNIYNKW/events.json","paper":"https://pith.science/paper/5FP2D5NO"},"agent_actions":{"view_html":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW","download_json":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW.json","view_paper":"https://pith.science/paper/5FP2D5NO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.09783&json=true","fetch_graph":"https://pith.science/api/pith-number/5FP2D5NO6JDBKAC4XRRYNIYNKW/graph.json","fetch_events":"https://pith.science/api/pith-number/5FP2D5NO6JDBKAC4XRRYNIYNKW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW/action/storage_attestation","attest_author":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW/action/author_attestation","sign_citation":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW/action/citation_signature","submit_replication":"https://pith.science/pith/5FP2D5NO6JDBKAC4XRRYNIYNKW/action/replication_record"}},"created_at":"2026-07-05T08:04:16.024535+00:00","updated_at":"2026-07-05T08:04:16.024535+00:00"}