{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:32NFY22FY52USPXXMYDOR6GR3E","short_pith_number":"pith:32NFY22F","schema_version":"1.0","canonical_sha256":"de9a5c6b45c775493ef76606e8f8d1d934b4a4a8c6036b61b2d1a8e9556feac3","source":{"kind":"arxiv","id":"2308.08493","version":3},"attestation_state":"computed","paper":{"title":"Time Travel in LLMs: Tracing Data Contamination in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Mihai Surdeanu, Shahriar Golchin","submitted_at":"2023-08-16T16:48:57Z","abstract_excerpt":"Data contamination, i.e., the presence of test data from downstream tasks in the training data of large language models (LLMs), is a potential major issue in measuring LLMs' real effectiveness on other tasks. We propose a straightforward yet effective method for identifying data contamination within LLMs. At its core, our approach starts by identifying potential contamination at the instance level; using this information, our approach then assesses wider contamination at the partition level. To estimate contamination of individual instances, we employ \"guided instruction:\" a prompt consisting "},"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":"2308.08493","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-08-16T16:48:57Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"5a6829efe222142752a94fceb73bd5e46c3b5b452b766ec6ddc2264d54fbac04","abstract_canon_sha256":"d018a10666ae5b07df4bd965037ea2ba82f55aad5a05527685e1e01b3b943123"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:47:58.442777Z","signature_b64":"Ab6Ac+A9pHDPIjd7jdTW7L3ccfQtW0VQvwbyT5I6egj8zQxugEcZ8/8pDQ5uHgvgymeFc/dNAUMXzW5rZsPjDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"de9a5c6b45c775493ef76606e8f8d1d934b4a4a8c6036b61b2d1a8e9556feac3","last_reissued_at":"2026-07-05T07:47:58.442200Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:47:58.442200Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Time Travel in LLMs: Tracing Data Contamination in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Mihai Surdeanu, Shahriar Golchin","submitted_at":"2023-08-16T16:48:57Z","abstract_excerpt":"Data contamination, i.e., the presence of test data from downstream tasks in the training data of large language models (LLMs), is a potential major issue in measuring LLMs' real effectiveness on other tasks. We propose a straightforward yet effective method for identifying data contamination within LLMs. At its core, our approach starts by identifying potential contamination at the instance level; using this information, our approach then assesses wider contamination at the partition level. To estimate contamination of individual instances, we employ \"guided instruction:\" a prompt consisting "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2308.08493","kind":"arxiv","version":3},"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/2308.08493/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":"2308.08493","created_at":"2026-07-05T07:47:58.442282+00:00"},{"alias_kind":"arxiv_version","alias_value":"2308.08493v3","created_at":"2026-07-05T07:47:58.442282+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.08493","created_at":"2026-07-05T07:47:58.442282+00:00"},{"alias_kind":"pith_short_12","alias_value":"32NFY22FY52U","created_at":"2026-07-05T07:47:58.442282+00:00"},{"alias_kind":"pith_short_16","alias_value":"32NFY22FY52USPXX","created_at":"2026-07-05T07:47:58.442282+00:00"},{"alias_kind":"pith_short_8","alias_value":"32NFY22F","created_at":"2026-07-05T07:47:58.442282+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.19224","citing_title":"iCoRe: An Iterative Correlation-Aware Retriever for Bug Reproduction Test Generation","ref_index":15,"is_internal_anchor":true},{"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":13,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01686","citing_title":"WARP: Weight-Space Analysis for Recovering Training Data Portfolios","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00062","citing_title":"AlgoBench: Benchmarking Algorithmic Adaptation in Code Generation","ref_index":8,"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":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16215","citing_title":"Fully Open Meditron: An Auditable Pipeline for Clinical LLMs","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29196","citing_title":"Representational Depth of Evaluation Awareness Shifts With Scale in Open-Weight Language Models","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2402.01411","citing_title":"CodePori: Large-Scale System for Autonomous Software Development Using Multi-Agent Technology","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21856","citing_title":"The Illusion of Reasoning: Exposing Evasive Data Contamination in LLMs via Zero-CoT Truncation","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16215","citing_title":"Fully Open Meditron: An Auditable Pipeline for Clinical LLMs","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17829","citing_title":"Interactive Evaluation Requires a Design Science","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17830","citing_title":"Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2506.20941","citing_title":"Revisiting the Past: Data Unlearning with Model State History","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2310.16789","citing_title":"Detecting Pretraining Data from Large Language Models","ref_index":113,"is_internal_anchor":false},{"citing_arxiv_id":"2404.11584","citing_title":"The Landscape of Emerging AI Agent Architectures for Reasoning, Planning, and Tool Calling: A Survey","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2310.11511","citing_title":"Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection","ref_index":167,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19224","citing_title":"iCoRe: An Iterative Correlation-Aware Retriever for Bug Reproduction Test Generation","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2403.07974","citing_title":"LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code","ref_index":254,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E","json":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E.json","graph_json":"https://pith.science/api/pith-number/32NFY22FY52USPXXMYDOR6GR3E/graph.json","events_json":"https://pith.science/api/pith-number/32NFY22FY52USPXXMYDOR6GR3E/events.json","paper":"https://pith.science/paper/32NFY22F"},"agent_actions":{"view_html":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E","download_json":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E.json","view_paper":"https://pith.science/paper/32NFY22F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2308.08493&json=true","fetch_graph":"https://pith.science/api/pith-number/32NFY22FY52USPXXMYDOR6GR3E/graph.json","fetch_events":"https://pith.science/api/pith-number/32NFY22FY52USPXXMYDOR6GR3E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E/action/storage_attestation","attest_author":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E/action/author_attestation","sign_citation":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E/action/citation_signature","submit_replication":"https://pith.science/pith/32NFY22FY52USPXXMYDOR6GR3E/action/replication_record"}},"created_at":"2026-07-05T07:47:58.442282+00:00","updated_at":"2026-07-05T07:47:58.442282+00:00"}