{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:73UFGPNJJK76ACDAFCENSSFTBS","short_pith_number":"pith:73UFGPNJ","schema_version":"1.0","canonical_sha256":"fee8533da94abfe008602888d948b30ca3a96fe13c2fbb254683436ec8a5a348","source":{"kind":"arxiv","id":"2208.01561","version":1},"attestation_state":"computed","paper":{"title":"Lost in Space Marking","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Cassandra L. Jacobs, Yuval Pinter","submitted_at":"2022-08-02T16:07:31Z","abstract_excerpt":"We look at a decision taken early in training a subword tokenizer, namely whether it should be the word-initial token that carries a special mark, or the word-final one. Based on surface-level considerations of efficiency and cohesion, as well as morphological coverage, we find that a Unigram LM tokenizer trained on pre-tokenized English text is better off marking the word-initial token, while one trained on raw text benefits from marking word ends. Our findings generalize across domains."},"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":"2208.01561","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-08-02T16:07:31Z","cross_cats_sorted":[],"title_canon_sha256":"5b1719988a8927a3b1aeaabf4c78f3a7faa312d3687835be82ab31a0045a15c5","abstract_canon_sha256":"d628db9e90f5789530e5b1337a6b60ccd36690dff45cf67889cfbd636412e99f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:45:35.036900Z","signature_b64":"KxhW3ic4Mhq4GxMrLAjQl4FB9pU6ps05wBYsZd6NXYyxgJbiGGWZ+uWTDgqLCzVmODbn7Xh6gtTWnD2+XmkkBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fee8533da94abfe008602888d948b30ca3a96fe13c2fbb254683436ec8a5a348","last_reissued_at":"2026-07-05T04:45:35.036492Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:45:35.036492Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Lost in Space Marking","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Cassandra L. Jacobs, Yuval Pinter","submitted_at":"2022-08-02T16:07:31Z","abstract_excerpt":"We look at a decision taken early in training a subword tokenizer, namely whether it should be the word-initial token that carries a special mark, or the word-final one. Based on surface-level considerations of efficiency and cohesion, as well as morphological coverage, we find that a Unigram LM tokenizer trained on pre-tokenized English text is better off marking the word-initial token, while one trained on raw text benefits from marking word ends. Our findings generalize across domains."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2208.01561","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/2208.01561/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":"2208.01561","created_at":"2026-07-05T04:45:35.036559+00:00"},{"alias_kind":"arxiv_version","alias_value":"2208.01561v1","created_at":"2026-07-05T04:45:35.036559+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2208.01561","created_at":"2026-07-05T04:45:35.036559+00:00"},{"alias_kind":"pith_short_12","alias_value":"73UFGPNJJK76","created_at":"2026-07-05T04:45:35.036559+00:00"},{"alias_kind":"pith_short_16","alias_value":"73UFGPNJJK76ACDA","created_at":"2026-07-05T04:45:35.036559+00:00"},{"alias_kind":"pith_short_8","alias_value":"73UFGPNJ","created_at":"2026-07-05T04:45:35.036559+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.18639","citing_title":"ByteSpan: Information-Driven Subword Tokenisation","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS","json":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS.json","graph_json":"https://pith.science/api/pith-number/73UFGPNJJK76ACDAFCENSSFTBS/graph.json","events_json":"https://pith.science/api/pith-number/73UFGPNJJK76ACDAFCENSSFTBS/events.json","paper":"https://pith.science/paper/73UFGPNJ"},"agent_actions":{"view_html":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS","download_json":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS.json","view_paper":"https://pith.science/paper/73UFGPNJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2208.01561&json=true","fetch_graph":"https://pith.science/api/pith-number/73UFGPNJJK76ACDAFCENSSFTBS/graph.json","fetch_events":"https://pith.science/api/pith-number/73UFGPNJJK76ACDAFCENSSFTBS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS/action/storage_attestation","attest_author":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS/action/author_attestation","sign_citation":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS/action/citation_signature","submit_replication":"https://pith.science/pith/73UFGPNJJK76ACDAFCENSSFTBS/action/replication_record"}},"created_at":"2026-07-05T04:45:35.036559+00:00","updated_at":"2026-07-05T04:45:35.036559+00:00"}