{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:G3RRHNIE6PMKI22HIWLOYOZIIQ","short_pith_number":"pith:G3RRHNIE","canonical_record":{"source":{"id":"2510.05699","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-07T09:05:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"219e7e7e5b54d3feec4c29b7433c2e6f760b38c518254d12a78be5486656f6c5","abstract_canon_sha256":"838e963313a9eb44b2cbf14bb5af28b2c8f4b4354d2be47c9ffedfa631be8717"},"schema_version":"1.0"},"canonical_sha256":"36e313b504f3d8a46b474596ec3b2844287cfa4c28d526b548cb24005b93f1c8","source":{"kind":"arxiv","id":"2510.05699","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.05699","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"arxiv_version","alias_value":"2510.05699v4","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.05699","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"pith_short_12","alias_value":"G3RRHNIE6PMK","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"pith_short_16","alias_value":"G3RRHNIE6PMKI22H","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"pith_short_8","alias_value":"G3RRHNIE","created_at":"2026-05-26T02:03:57Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:G3RRHNIE6PMKI22HIWLOYOZIIQ","target":"record","payload":{"canonical_record":{"source":{"id":"2510.05699","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-07T09:05:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"219e7e7e5b54d3feec4c29b7433c2e6f760b38c518254d12a78be5486656f6c5","abstract_canon_sha256":"838e963313a9eb44b2cbf14bb5af28b2c8f4b4354d2be47c9ffedfa631be8717"},"schema_version":"1.0"},"canonical_sha256":"36e313b504f3d8a46b474596ec3b2844287cfa4c28d526b548cb24005b93f1c8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:03:57.140583Z","signature_b64":"Qlia+Ox1sqEWgR68IFkJSURNxYcx6o4DLt+YWNsDn4D+Xa7Ka0ulXsfSTrtm3aVNc0ibzk4FPi85OEC4Ozn/AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"36e313b504f3d8a46b474596ec3b2844287cfa4c28d526b548cb24005b93f1c8","last_reissued_at":"2026-05-26T02:03:57.139626Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:03:57.139626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2510.05699","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T02:03:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZdvKCcUtuIj1Uf3GbeS+JL2evWfztp9uFxQxcM4jMNQ2tmXT6oKo1kwCaK0KhXU3BlGCuX34PHISDi+yTDy0Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T05:08:01.057597Z"},"content_sha256":"58751fdb96ebacb4e4483099b35dc30c144110e31a2bfbb23ac888f7d5b11d5d","schema_version":"1.0","event_id":"sha256:58751fdb96ebacb4e4483099b35dc30c144110e31a2bfbb23ac888f7d5b11d5d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:G3RRHNIE6PMKI22HIWLOYOZIIQ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Membership Inference Attacks on Tokenizers of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Kejiang Chen, Meng Tong, Ninghui Li, Weiming Zhang, Yuntao Du","submitted_at":"2025-10-07T09:05:40Z","abstract_excerpt":"Membership inference attacks (MIAs) are widely used to assess the privacy risks associated with machine learning models. However, when these attacks are applied to pre-trained large language models (LLMs), they encounter significant challenges, including mislabeled samples, distribution shifts, and discrepancies in model size between experimental and real-world settings. To address these limitations, we introduce tokenizers as a new attack vector for membership inference. Specifically, a tokenizer converts raw text into tokens for LLMs. Unlike full models, tokenizers can be efficiently trained"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.05699","kind":"arxiv","version":4},"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/2510.05699/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T02:03:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kMqM6KYUQcWBZW/sOvrVyOYNF993A498wu31tAR9mjB0BvHXR2JWi5CESG9ntoEbdbzSBC49eqLGk+VJreONAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T05:08:01.058362Z"},"content_sha256":"fb22086610e274bd16a5efc4d6397a122a96b2439690a27e8294e69dde133715","schema_version":"1.0","event_id":"sha256:fb22086610e274bd16a5efc4d6397a122a96b2439690a27e8294e69dde133715"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/G3RRHNIE6PMKI22HIWLOYOZIIQ/bundle.json","state_url":"https://pith.science/pith/G3RRHNIE6PMKI22HIWLOYOZIIQ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/G3RRHNIE6PMKI22HIWLOYOZIIQ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-27T05:08:01Z","links":{"resolver":"https://pith.science/pith/G3RRHNIE6PMKI22HIWLOYOZIIQ","bundle":"https://pith.science/pith/G3RRHNIE6PMKI22HIWLOYOZIIQ/bundle.json","state":"https://pith.science/pith/G3RRHNIE6PMKI22HIWLOYOZIIQ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/G3RRHNIE6PMKI22HIWLOYOZIIQ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:G3RRHNIE6PMKI22HIWLOYOZIIQ","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"838e963313a9eb44b2cbf14bb5af28b2c8f4b4354d2be47c9ffedfa631be8717","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-07T09:05:40Z","title_canon_sha256":"219e7e7e5b54d3feec4c29b7433c2e6f760b38c518254d12a78be5486656f6c5"},"schema_version":"1.0","source":{"id":"2510.05699","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.05699","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"arxiv_version","alias_value":"2510.05699v4","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.05699","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"pith_short_12","alias_value":"G3RRHNIE6PMK","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"pith_short_16","alias_value":"G3RRHNIE6PMKI22H","created_at":"2026-05-26T02:03:57Z"},{"alias_kind":"pith_short_8","alias_value":"G3RRHNIE","created_at":"2026-05-26T02:03:57Z"}],"graph_snapshots":[{"event_id":"sha256:fb22086610e274bd16a5efc4d6397a122a96b2439690a27e8294e69dde133715","target":"graph","created_at":"2026-05-26T02:03:57Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2510.05699/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Membership inference attacks (MIAs) are widely used to assess the privacy risks associated with machine learning models. However, when these attacks are applied to pre-trained large language models (LLMs), they encounter significant challenges, including mislabeled samples, distribution shifts, and discrepancies in model size between experimental and real-world settings. To address these limitations, we introduce tokenizers as a new attack vector for membership inference. Specifically, a tokenizer converts raw text into tokens for LLMs. Unlike full models, tokenizers can be efficiently trained","authors_text":"Kejiang Chen, Meng Tong, Ninghui Li, Weiming Zhang, Yuntao Du","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-07T09:05:40Z","title":"Membership Inference Attacks on Tokenizers of Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.05699","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:58751fdb96ebacb4e4483099b35dc30c144110e31a2bfbb23ac888f7d5b11d5d","target":"record","created_at":"2026-05-26T02:03:57Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"838e963313a9eb44b2cbf14bb5af28b2c8f4b4354d2be47c9ffedfa631be8717","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-07T09:05:40Z","title_canon_sha256":"219e7e7e5b54d3feec4c29b7433c2e6f760b38c518254d12a78be5486656f6c5"},"schema_version":"1.0","source":{"id":"2510.05699","kind":"arxiv","version":4}},"canonical_sha256":"36e313b504f3d8a46b474596ec3b2844287cfa4c28d526b548cb24005b93f1c8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"36e313b504f3d8a46b474596ec3b2844287cfa4c28d526b548cb24005b93f1c8","first_computed_at":"2026-05-26T02:03:57.139626Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T02:03:57.139626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Qlia+Ox1sqEWgR68IFkJSURNxYcx6o4DLt+YWNsDn4D+Xa7Ka0ulXsfSTrtm3aVNc0ibzk4FPi85OEC4Ozn/AQ==","signature_status":"signed_v1","signed_at":"2026-05-26T02:03:57.140583Z","signed_message":"canonical_sha256_bytes"},"source_id":"2510.05699","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:58751fdb96ebacb4e4483099b35dc30c144110e31a2bfbb23ac888f7d5b11d5d","sha256:fb22086610e274bd16a5efc4d6397a122a96b2439690a27e8294e69dde133715"],"state_sha256":"f2a7749163a98dcd904f378040c9d8a7d2fcd8f18ddc0c724f63a9eae6bbb739"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bmthpNzXKQ9h8woufd1OW0q0rcPV/TSGUvL2VVVB5XlfVUo90tstDBB1Ar44G5tC8SM3LpGzP3RpkDOwM7XIAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-27T05:08:01.062557Z","bundle_sha256":"56f1d465b1c9e36240edb52bea7d41e5b5aa58aee0637c609f09f664a41bbfb1"}}