{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JF66CJ74JJVBUZI7CMYVC3DRA7","short_pith_number":"pith:JF66CJ74","schema_version":"1.0","canonical_sha256":"497de127fc4a6a1a651f1331516c7107c04fb7337e9ae9994e1a550f885f1a6f","source":{"kind":"arxiv","id":"2508.06621","version":1},"attestation_state":"computed","paper":{"title":"Train It and Forget It: Merge Lists are Unnecessary for BPE Inference in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kartik Goyal, Tomohiro Sawada","submitted_at":"2025-08-08T18:10:03Z","abstract_excerpt":"Standard Byte-Pair Encoding (BPE) tokenization compresses text by pairing a learned token vocabulary with a detailed merge list. Recent work has shown that this merge list exposes a potential attack surface for extracting information about language model's training data. In this paper, we explore the downstream impact of BPE inference algorithms that do not rely on this merge list at all, and hence differ from the encoding process during BPE training. To address this question, we investigate two broad classes of BPE inference schemes that differ from BPE application during training: a) targete"},"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":"2508.06621","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-08T18:10:03Z","cross_cats_sorted":[],"title_canon_sha256":"f8f55e9276f9cb774334c35f0054bbadebcad618e42c3c24613b72c049c9434a","abstract_canon_sha256":"c26c76d7289f1637589d4c80d5803185e392ea76957d4aa17062d8cd22f31478"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:51:03.675127Z","signature_b64":"Z5buGxgmgmVAn/zinLlfivo9F4VrYEhncqjPhSbr1N3o1NpIlihXbMZAIyjbSmCxGQP6nN+2iLIrKthunA7TDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"497de127fc4a6a1a651f1331516c7107c04fb7337e9ae9994e1a550f885f1a6f","last_reissued_at":"2026-07-05T11:51:03.674631Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:51:03.674631Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Train It and Forget It: Merge Lists are Unnecessary for BPE Inference in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kartik Goyal, Tomohiro Sawada","submitted_at":"2025-08-08T18:10:03Z","abstract_excerpt":"Standard Byte-Pair Encoding (BPE) tokenization compresses text by pairing a learned token vocabulary with a detailed merge list. Recent work has shown that this merge list exposes a potential attack surface for extracting information about language model's training data. In this paper, we explore the downstream impact of BPE inference algorithms that do not rely on this merge list at all, and hence differ from the encoding process during BPE training. To address this question, we investigate two broad classes of BPE inference schemes that differ from BPE application during training: a) targete"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.06621","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/2508.06621/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":"2508.06621","created_at":"2026-07-05T11:51:03.674694+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.06621v1","created_at":"2026-07-05T11:51:03.674694+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.06621","created_at":"2026-07-05T11:51:03.674694+00:00"},{"alias_kind":"pith_short_12","alias_value":"JF66CJ74JJVB","created_at":"2026-07-05T11:51:03.674694+00:00"},{"alias_kind":"pith_short_16","alias_value":"JF66CJ74JJVBUZI7","created_at":"2026-07-05T11:51:03.674694+00:00"},{"alias_kind":"pith_short_8","alias_value":"JF66CJ74","created_at":"2026-07-05T11:51:03.674694+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/JF66CJ74JJVBUZI7CMYVC3DRA7","json":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7.json","graph_json":"https://pith.science/api/pith-number/JF66CJ74JJVBUZI7CMYVC3DRA7/graph.json","events_json":"https://pith.science/api/pith-number/JF66CJ74JJVBUZI7CMYVC3DRA7/events.json","paper":"https://pith.science/paper/JF66CJ74"},"agent_actions":{"view_html":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7","download_json":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7.json","view_paper":"https://pith.science/paper/JF66CJ74","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.06621&json=true","fetch_graph":"https://pith.science/api/pith-number/JF66CJ74JJVBUZI7CMYVC3DRA7/graph.json","fetch_events":"https://pith.science/api/pith-number/JF66CJ74JJVBUZI7CMYVC3DRA7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7/action/storage_attestation","attest_author":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7/action/author_attestation","sign_citation":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7/action/citation_signature","submit_replication":"https://pith.science/pith/JF66CJ74JJVBUZI7CMYVC3DRA7/action/replication_record"}},"created_at":"2026-07-05T11:51:03.674694+00:00","updated_at":"2026-07-05T11:51:03.674694+00:00"}