{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:5QGGTNODICGZQ67O7FDJIEDYPJ","short_pith_number":"pith:5QGGTNOD","canonical_record":{"source":{"id":"2407.08818","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T18:59:21Z","cross_cats_sorted":[],"title_canon_sha256":"668b8900bb0ff0b84f9d7c4f19982f4089c1e4d954200b772161c37a44932502","abstract_canon_sha256":"d2e9e7d0182b296a05d3b51f5cd71e25a30be98a8db20b80795bd9459539f58a"},"schema_version":"1.0"},"canonical_sha256":"ec0c69b5c3408d987beef9469410787a500b6bba2e6e83cef56105d35e598596","source":{"kind":"arxiv","id":"2407.08818","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.08818","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"arxiv_version","alias_value":"2407.08818v2","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.08818","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"pith_short_12","alias_value":"5QGGTNODICGZ","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"pith_short_16","alias_value":"5QGGTNODICGZQ67O","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"pith_short_8","alias_value":"5QGGTNOD","created_at":"2026-07-05T09:36:13Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:5QGGTNODICGZQ67O7FDJIEDYPJ","target":"record","payload":{"canonical_record":{"source":{"id":"2407.08818","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T18:59:21Z","cross_cats_sorted":[],"title_canon_sha256":"668b8900bb0ff0b84f9d7c4f19982f4089c1e4d954200b772161c37a44932502","abstract_canon_sha256":"d2e9e7d0182b296a05d3b51f5cd71e25a30be98a8db20b80795bd9459539f58a"},"schema_version":"1.0"},"canonical_sha256":"ec0c69b5c3408d987beef9469410787a500b6bba2e6e83cef56105d35e598596","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:36:13.089365Z","signature_b64":"4VrCAXDo189uMCrpmizQ73SeNnWHv0MFEDhIcaflIvgPDz7+jM1GocBHoGxw4rNXzM4iadl0L0nt5cCn8SAwCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ec0c69b5c3408d987beef9469410787a500b6bba2e6e83cef56105d35e598596","last_reissued_at":"2026-07-05T09:36:13.088911Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:36:13.088911Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2407.08818","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:36:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kpqA5MvNO7RL+Vr+gtlhAlbVClQWuNvUoCZq0jQQ9pZGn7n2DwHnXWMDhkovumOBn0rFEmIv7+WI/cqttXjVDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T03:39:07.940545Z"},"content_sha256":"22aa59199a142705c7f11384dc98dcd744b83a2e9cafb32e2d86cdef4b6f7f6e","schema_version":"1.0","event_id":"sha256:22aa59199a142705c7f11384dc98dcd744b83a2e9cafb32e2d86cdef4b6f7f6e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:5QGGTNODICGZQ67O7FDJIEDYPJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"MAGNET: Improving the Multilingual Fairness of Language Models with Adaptive Gradient-Based Tokenization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hila Gonen, Noah A. Smith, Orevaoghene Ahia, Sachin Kumar, Tomasz Limisiewicz, Valentin Hofmann, Yulia Tsvetkov","submitted_at":"2024-07-11T18:59:21Z","abstract_excerpt":"In multilingual settings, non-Latin scripts and low-resource languages are usually disadvantaged in terms of language models' utility, efficiency, and cost. Specifically, previous studies have reported multiple modeling biases that the current tokenization algorithms introduce to non-Latin script languages, the main one being over-segmentation. In this work, we propose MAGNET; multilingual adaptive gradient-based tokenization to reduce over-segmentation via adaptive gradient-based subword tokenization. MAGNET learns to predict segment boundaries between byte tokens in a sequence via sub-module"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.08818","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/2407.08818/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":"2026-07-05T09:36:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rpIBPtVFL+Rb4FNlJ7aoMsYHoTsnWwHtPwCw4WKtdqsqEK0OheBT+wo0z93FFaKyHjwbQWCjztdubtdsm4o/Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T03:39:07.941057Z"},"content_sha256":"d6b5d5cb50313dff1f198f9d65a8410015f4d87975e12e83a3fe1a31ed4cbb0d","schema_version":"1.0","event_id":"sha256:d6b5d5cb50313dff1f198f9d65a8410015f4d87975e12e83a3fe1a31ed4cbb0d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5QGGTNODICGZQ67O7FDJIEDYPJ/bundle.json","state_url":"https://pith.science/pith/5QGGTNODICGZQ67O7FDJIEDYPJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5QGGTNODICGZQ67O7FDJIEDYPJ/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":"2026-08-07T03:39:07Z","links":{"resolver":"https://pith.science/pith/5QGGTNODICGZQ67O7FDJIEDYPJ","bundle":"https://pith.science/pith/5QGGTNODICGZQ67O7FDJIEDYPJ/bundle.json","state":"https://pith.science/pith/5QGGTNODICGZQ67O7FDJIEDYPJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5QGGTNODICGZQ67O7FDJIEDYPJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:5QGGTNODICGZQ67O7FDJIEDYPJ","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":"d2e9e7d0182b296a05d3b51f5cd71e25a30be98a8db20b80795bd9459539f58a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T18:59:21Z","title_canon_sha256":"668b8900bb0ff0b84f9d7c4f19982f4089c1e4d954200b772161c37a44932502"},"schema_version":"1.0","source":{"id":"2407.08818","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.08818","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"arxiv_version","alias_value":"2407.08818v2","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.08818","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"pith_short_12","alias_value":"5QGGTNODICGZ","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"pith_short_16","alias_value":"5QGGTNODICGZQ67O","created_at":"2026-07-05T09:36:13Z"},{"alias_kind":"pith_short_8","alias_value":"5QGGTNOD","created_at":"2026-07-05T09:36:13Z"}],"graph_snapshots":[{"event_id":"sha256:d6b5d5cb50313dff1f198f9d65a8410015f4d87975e12e83a3fe1a31ed4cbb0d","target":"graph","created_at":"2026-07-05T09:36:13Z","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/2407.08818/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In multilingual settings, non-Latin scripts and low-resource languages are usually disadvantaged in terms of language models' utility, efficiency, and cost. Specifically, previous studies have reported multiple modeling biases that the current tokenization algorithms introduce to non-Latin script languages, the main one being over-segmentation. In this work, we propose MAGNET; multilingual adaptive gradient-based tokenization to reduce over-segmentation via adaptive gradient-based subword tokenization. MAGNET learns to predict segment boundaries between byte tokens in a sequence via sub-module","authors_text":"Hila Gonen, Noah A. Smith, Orevaoghene Ahia, Sachin Kumar, Tomasz Limisiewicz, Valentin Hofmann, Yulia Tsvetkov","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T18:59:21Z","title":"MAGNET: Improving the Multilingual Fairness of Language Models with Adaptive Gradient-Based Tokenization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.08818","kind":"arxiv","version":2},"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":"sha256:22aa59199a142705c7f11384dc98dcd744b83a2e9cafb32e2d86cdef4b6f7f6e","target":"record","created_at":"2026-07-05T09:36:13Z","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":"d2e9e7d0182b296a05d3b51f5cd71e25a30be98a8db20b80795bd9459539f58a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T18:59:21Z","title_canon_sha256":"668b8900bb0ff0b84f9d7c4f19982f4089c1e4d954200b772161c37a44932502"},"schema_version":"1.0","source":{"id":"2407.08818","kind":"arxiv","version":2}},"canonical_sha256":"ec0c69b5c3408d987beef9469410787a500b6bba2e6e83cef56105d35e598596","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ec0c69b5c3408d987beef9469410787a500b6bba2e6e83cef56105d35e598596","first_computed_at":"2026-07-05T09:36:13.088911Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:36:13.088911Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"4VrCAXDo189uMCrpmizQ73SeNnWHv0MFEDhIcaflIvgPDz7+jM1GocBHoGxw4rNXzM4iadl0L0nt5cCn8SAwCA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:36:13.089365Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.08818","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:22aa59199a142705c7f11384dc98dcd744b83a2e9cafb32e2d86cdef4b6f7f6e","sha256:d6b5d5cb50313dff1f198f9d65a8410015f4d87975e12e83a3fe1a31ed4cbb0d"],"state_sha256":"db9e977ba30b64c9e79ac04d5233b984e3fcb19f35702d20fef3375df056b50c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XlvjpCny6bk+hUHvMlCAoMu+AJ/O+DwrBMqCJ+FO/yvjZ9KRp5wMem5k13jOjkMJRuLGW/BW15nP5a+J8XPMBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T03:39:07.945718Z","bundle_sha256":"aa04c5b46fbd1e1c3feb34b94e86068180f701fc10cdc519130e0f1ecd608077"}}