{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:RS4K2CIL6OHHSK4QANYWNIDPYX","short_pith_number":"pith:RS4K2CIL","schema_version":"1.0","canonical_sha256":"8cb8ad090bf38e792b90037166a06fc5c6f4b081260c0e709ff7bc20b2dc4833","source":{"kind":"arxiv","id":"2505.24689","version":1},"attestation_state":"computed","paper":{"title":"BPE Stays on SCRIPT: Structured Encoding for Robust Multilingual Pretokenization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Catherine Arnett, Sander Land","submitted_at":"2025-05-30T15:12:41Z","abstract_excerpt":"Byte Pair Encoding (BPE) tokenizers, widely used in Large Language Models, face challenges in multilingual settings, including penalization of non-Western scripts and the creation of tokens with partial UTF-8 sequences. Pretokenization, often reliant on complex regular expressions, can also introduce fragility and unexpected edge cases. We propose SCRIPT (Script Category Representation in PreTokenization), a novel encoding scheme that bypasses UTF-8 byte conversion by using initial tokens based on Unicode script and category properties. This approach enables a simple, rule-based pretokenizatio"},"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":"2505.24689","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T15:12:41Z","cross_cats_sorted":[],"title_canon_sha256":"6bec071f28294bbf0473c1639ec4f87a5175f6020299f0768f8ed2c600450a8b","abstract_canon_sha256":"a884b3b190531d2b30482c04230f4521a2445f80a76c97733f72c2579a33c572"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:54.741974Z","signature_b64":"GdOtc5eSpFNQUYD7j5Try+zS/oydGJ1i3TRn9tk93IlubaZsCpePuzf8G/FeTXMh1HqGv1YRmYag4n8n1ddbBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8cb8ad090bf38e792b90037166a06fc5c6f4b081260c0e709ff7bc20b2dc4833","last_reissued_at":"2026-07-05T11:12:54.741391Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:54.741391Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"BPE Stays on SCRIPT: Structured Encoding for Robust Multilingual Pretokenization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Catherine Arnett, Sander Land","submitted_at":"2025-05-30T15:12:41Z","abstract_excerpt":"Byte Pair Encoding (BPE) tokenizers, widely used in Large Language Models, face challenges in multilingual settings, including penalization of non-Western scripts and the creation of tokens with partial UTF-8 sequences. Pretokenization, often reliant on complex regular expressions, can also introduce fragility and unexpected edge cases. We propose SCRIPT (Script Category Representation in PreTokenization), a novel encoding scheme that bypasses UTF-8 byte conversion by using initial tokens based on Unicode script and category properties. This approach enables a simple, rule-based pretokenizatio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.24689","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/2505.24689/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":"2505.24689","created_at":"2026-07-05T11:12:54.741455+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.24689v1","created_at":"2026-07-05T11:12:54.741455+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.24689","created_at":"2026-07-05T11:12:54.741455+00:00"},{"alias_kind":"pith_short_12","alias_value":"RS4K2CIL6OHH","created_at":"2026-07-05T11:12:54.741455+00:00"},{"alias_kind":"pith_short_16","alias_value":"RS4K2CIL6OHHSK4Q","created_at":"2026-07-05T11:12:54.741455+00:00"},{"alias_kind":"pith_short_8","alias_value":"RS4K2CIL","created_at":"2026-07-05T11:12:54.741455+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27019","citing_title":"MinGram: A Minimalist Unigram Tokenizer with High Compression and Competitive Morphological Alignment","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22705","citing_title":"Tokenization with Split Trees","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2601.07220","citing_title":"The Roots of Performance Disparity in Multilingual Language Models: Intrinsic Modeling Difficulty or Design Choices?","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05192","citing_title":"Faster Superword Tokenization","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX","json":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX.json","graph_json":"https://pith.science/api/pith-number/RS4K2CIL6OHHSK4QANYWNIDPYX/graph.json","events_json":"https://pith.science/api/pith-number/RS4K2CIL6OHHSK4QANYWNIDPYX/events.json","paper":"https://pith.science/paper/RS4K2CIL"},"agent_actions":{"view_html":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX","download_json":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX.json","view_paper":"https://pith.science/paper/RS4K2CIL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.24689&json=true","fetch_graph":"https://pith.science/api/pith-number/RS4K2CIL6OHHSK4QANYWNIDPYX/graph.json","fetch_events":"https://pith.science/api/pith-number/RS4K2CIL6OHHSK4QANYWNIDPYX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX/action/storage_attestation","attest_author":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX/action/author_attestation","sign_citation":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX/action/citation_signature","submit_replication":"https://pith.science/pith/RS4K2CIL6OHHSK4QANYWNIDPYX/action/replication_record"}},"created_at":"2026-07-05T11:12:54.741455+00:00","updated_at":"2026-07-05T11:12:54.741455+00:00"}