{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BLXYOMCW6SR4S54F25Z4JQMRHP","short_pith_number":"pith:BLXYOMCW","schema_version":"1.0","canonical_sha256":"0aef873056f4a3c97785d773c4c1913be3a9a74af797d18a77ac7ad98990f4c7","source":{"kind":"arxiv","id":"2405.05417","version":2},"attestation_state":"computed","paper":{"title":"Fishing for Magikarp: Automatically Detecting Under-trained Tokens in Large Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Max Bartolo, Sander Land","submitted_at":"2024-05-08T20:37:56Z","abstract_excerpt":"The disconnect between tokenizer creation and model training in language models allows for specific inputs, such as the infamous SolidGoldMagikarp token, to induce unwanted model behaviour. Although such `glitch tokens', tokens present in the tokenizer vocabulary but that are nearly or entirely absent during model training, have been observed across various models, a reliable method to identify and address them has been missing. We present a comprehensive analysis of Large Language Model tokenizers, specifically targeting this issue of detecting under-trained tokens. Through a combination of t"},"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":"2405.05417","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-08T20:37:56Z","cross_cats_sorted":[],"title_canon_sha256":"459f3cf4a99e5b9ada4deedac5af4b53eca8b99102bbe931040ab248dea00f72","abstract_canon_sha256":"80402f28dcef7c44744b3467e7d5a7aa56b0103f5466c3212808f5240d8ff5a1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:12:28.299044Z","signature_b64":"vdb0DfjZBiEUvm5vl/eCqDiNg+Rp5MEHbogDUvv59k4OkxHpV/uMm41eo25Jh3M1BhTN0YRe2HNrihP6hqSLDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0aef873056f4a3c97785d773c4c1913be3a9a74af797d18a77ac7ad98990f4c7","last_reissued_at":"2026-07-05T09:12:28.298504Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:12:28.298504Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fishing for Magikarp: Automatically Detecting Under-trained Tokens in Large Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Max Bartolo, Sander Land","submitted_at":"2024-05-08T20:37:56Z","abstract_excerpt":"The disconnect between tokenizer creation and model training in language models allows for specific inputs, such as the infamous SolidGoldMagikarp token, to induce unwanted model behaviour. Although such `glitch tokens', tokens present in the tokenizer vocabulary but that are nearly or entirely absent during model training, have been observed across various models, a reliable method to identify and address them has been missing. We present a comprehensive analysis of Large Language Model tokenizers, specifically targeting this issue of detecting under-trained tokens. Through a combination of t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.05417","kind":"arxiv","version":2},"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/2405.05417/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":"2405.05417","created_at":"2026-07-05T09:12:28.298563+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.05417v2","created_at":"2026-07-05T09:12:28.298563+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.05417","created_at":"2026-07-05T09:12:28.298563+00:00"},{"alias_kind":"pith_short_12","alias_value":"BLXYOMCW6SR4","created_at":"2026-07-05T09:12:28.298563+00:00"},{"alias_kind":"pith_short_16","alias_value":"BLXYOMCW6SR4S54F","created_at":"2026-07-05T09:12:28.298563+00:00"},{"alias_kind":"pith_short_8","alias_value":"BLXYOMCW","created_at":"2026-07-05T09:12:28.298563+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2502.10248","citing_title":"Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model","ref_index":276,"is_internal_anchor":false},{"citing_arxiv_id":"2507.12720","citing_title":"FLEXITOKENS: Flexible Tokenization for Evolving Language Models","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26355","citing_title":"Shorthand for Thought: Compressing LLM Reasoning via Entropy-Guided Supertokens","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09630","citing_title":"Scratchpad Patching: Decoupling Compute from Patch Size in Byte-Level Language Models","ref_index":55,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP","json":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP.json","graph_json":"https://pith.science/api/pith-number/BLXYOMCW6SR4S54F25Z4JQMRHP/graph.json","events_json":"https://pith.science/api/pith-number/BLXYOMCW6SR4S54F25Z4JQMRHP/events.json","paper":"https://pith.science/paper/BLXYOMCW"},"agent_actions":{"view_html":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP","download_json":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP.json","view_paper":"https://pith.science/paper/BLXYOMCW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.05417&json=true","fetch_graph":"https://pith.science/api/pith-number/BLXYOMCW6SR4S54F25Z4JQMRHP/graph.json","fetch_events":"https://pith.science/api/pith-number/BLXYOMCW6SR4S54F25Z4JQMRHP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP/action/storage_attestation","attest_author":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP/action/author_attestation","sign_citation":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP/action/citation_signature","submit_replication":"https://pith.science/pith/BLXYOMCW6SR4S54F25Z4JQMRHP/action/replication_record"}},"created_at":"2026-07-05T09:12:28.298563+00:00","updated_at":"2026-07-05T09:12:28.298563+00:00"}