{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:74XTPKQ7IZNAANZKTQ3XUROLU3","short_pith_number":"pith:74XTPKQ7","canonical_record":{"source":{"id":"2402.12030","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T10:37:29Z","cross_cats_sorted":[],"title_canon_sha256":"bd1884ec333971300fb80ce995bb0aca29843b7c9cbf3bb2bc2260e984e0eabb","abstract_canon_sha256":"2335558256e9a1cc1aa9d928b1e704c09554e4f3e483fadcd71d739af5084a21"},"schema_version":"1.0"},"canonical_sha256":"ff2f37aa1f465a00372a9c377a45cba6e0ef3169555ef6195ed1bfe48bd2c64f","source":{"kind":"arxiv","id":"2402.12030","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.12030","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"arxiv_version","alias_value":"2402.12030v3","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.12030","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"pith_short_12","alias_value":"74XTPKQ7IZNA","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"pith_short_16","alias_value":"74XTPKQ7IZNAANZK","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"pith_short_8","alias_value":"74XTPKQ7","created_at":"2026-07-05T10:05:41Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:74XTPKQ7IZNAANZKTQ3XUROLU3","target":"record","payload":{"canonical_record":{"source":{"id":"2402.12030","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T10:37:29Z","cross_cats_sorted":[],"title_canon_sha256":"bd1884ec333971300fb80ce995bb0aca29843b7c9cbf3bb2bc2260e984e0eabb","abstract_canon_sha256":"2335558256e9a1cc1aa9d928b1e704c09554e4f3e483fadcd71d739af5084a21"},"schema_version":"1.0"},"canonical_sha256":"ff2f37aa1f465a00372a9c377a45cba6e0ef3169555ef6195ed1bfe48bd2c64f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:05:41.668946Z","signature_b64":"OpD7EURQ0cr2UUJlGVP3MWEE9Ex9kDJ84IDCcoV/wyPz3SzVU5AU5oq0p+3RiqDC+wb5hjiztlqR+Z+cqjQLDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff2f37aa1f465a00372a9c377a45cba6e0ef3169555ef6195ed1bfe48bd2c64f","last_reissued_at":"2026-07-05T10:05:41.668414Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:05:41.668414Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.12030","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:05:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1pJ66b46GM+MmbLlh8pLgmlCk9nxwAVq3VekfaIphpRatXpSxMOZ6qKw7oQrpbs6+8QInyd7YK497wtQeZukDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T19:42:05.523409Z"},"content_sha256":"54d0da17423ac40091c872e341669531a0ff3fdba97a119cb0ffffe756a76f2f","schema_version":"1.0","event_id":"sha256:54d0da17423ac40091c872e341669531a0ff3fdba97a119cb0ffffe756a76f2f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:74XTPKQ7IZNAANZKTQ3XUROLU3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Towards Cross-Tokenizer Distillation: the Universal Logit Distillation Loss for LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"C\\'eline Hudelot, Kevin El Haddad, Nicolas Boizard, Pierre Colombo","submitted_at":"2024-02-19T10:37:29Z","abstract_excerpt":"Deploying large language models (LLMs) of several billion parameters can be impractical in most industrial use cases due to constraints such as cost, latency limitations, and hardware accessibility. Knowledge distillation (KD) offers a solution by compressing knowledge from resource-intensive large models to smaller ones. Various strategies exist, some relying on the text generated by the teacher model and optionally utilizing his logits to enhance learning. However, these methods based on logits often require both teacher and student models to share the same tokenizer, limiting their applicab"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.12030","kind":"arxiv","version":3},"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/2402.12030/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":"2026-07-05T10:05:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qTWDkKt62jb+kKHwedoXQ+7ozRw90Wd0eG9MU4oHjw23nhkdxTolE+HceRy26P3LIPj21dr2dd8N8EMPtW/mDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T19:42:05.524005Z"},"content_sha256":"ac25a537a251a334b6974f14b7d08ce54f0042057f9041ee1d98f76378329c76","schema_version":"1.0","event_id":"sha256:ac25a537a251a334b6974f14b7d08ce54f0042057f9041ee1d98f76378329c76"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/74XTPKQ7IZNAANZKTQ3XUROLU3/bundle.json","state_url":"https://pith.science/pith/74XTPKQ7IZNAANZKTQ3XUROLU3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/74XTPKQ7IZNAANZKTQ3XUROLU3/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":"2026-08-13T19:42:05Z","links":{"resolver":"https://pith.science/pith/74XTPKQ7IZNAANZKTQ3XUROLU3","bundle":"https://pith.science/pith/74XTPKQ7IZNAANZKTQ3XUROLU3/bundle.json","state":"https://pith.science/pith/74XTPKQ7IZNAANZKTQ3XUROLU3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/74XTPKQ7IZNAANZKTQ3XUROLU3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:74XTPKQ7IZNAANZKTQ3XUROLU3","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":"2335558256e9a1cc1aa9d928b1e704c09554e4f3e483fadcd71d739af5084a21","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T10:37:29Z","title_canon_sha256":"bd1884ec333971300fb80ce995bb0aca29843b7c9cbf3bb2bc2260e984e0eabb"},"schema_version":"1.0","source":{"id":"2402.12030","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.12030","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"arxiv_version","alias_value":"2402.12030v3","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.12030","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"pith_short_12","alias_value":"74XTPKQ7IZNA","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"pith_short_16","alias_value":"74XTPKQ7IZNAANZK","created_at":"2026-07-05T10:05:41Z"},{"alias_kind":"pith_short_8","alias_value":"74XTPKQ7","created_at":"2026-07-05T10:05:41Z"}],"graph_snapshots":[{"event_id":"sha256:ac25a537a251a334b6974f14b7d08ce54f0042057f9041ee1d98f76378329c76","target":"graph","created_at":"2026-07-05T10:05:41Z","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/2402.12030/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Deploying large language models (LLMs) of several billion parameters can be impractical in most industrial use cases due to constraints such as cost, latency limitations, and hardware accessibility. Knowledge distillation (KD) offers a solution by compressing knowledge from resource-intensive large models to smaller ones. Various strategies exist, some relying on the text generated by the teacher model and optionally utilizing his logits to enhance learning. However, these methods based on logits often require both teacher and student models to share the same tokenizer, limiting their applicab","authors_text":"C\\'eline Hudelot, Kevin El Haddad, Nicolas Boizard, Pierre Colombo","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T10:37:29Z","title":"Towards Cross-Tokenizer Distillation: the Universal Logit Distillation Loss for LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.12030","kind":"arxiv","version":3},"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":"sha256:54d0da17423ac40091c872e341669531a0ff3fdba97a119cb0ffffe756a76f2f","target":"record","created_at":"2026-07-05T10:05:41Z","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":"2335558256e9a1cc1aa9d928b1e704c09554e4f3e483fadcd71d739af5084a21","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T10:37:29Z","title_canon_sha256":"bd1884ec333971300fb80ce995bb0aca29843b7c9cbf3bb2bc2260e984e0eabb"},"schema_version":"1.0","source":{"id":"2402.12030","kind":"arxiv","version":3}},"canonical_sha256":"ff2f37aa1f465a00372a9c377a45cba6e0ef3169555ef6195ed1bfe48bd2c64f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ff2f37aa1f465a00372a9c377a45cba6e0ef3169555ef6195ed1bfe48bd2c64f","first_computed_at":"2026-07-05T10:05:41.668414Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:05:41.668414Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"OpD7EURQ0cr2UUJlGVP3MWEE9Ex9kDJ84IDCcoV/wyPz3SzVU5AU5oq0p+3RiqDC+wb5hjiztlqR+Z+cqjQLDA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:05:41.668946Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.12030","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:54d0da17423ac40091c872e341669531a0ff3fdba97a119cb0ffffe756a76f2f","sha256:ac25a537a251a334b6974f14b7d08ce54f0042057f9041ee1d98f76378329c76"],"state_sha256":"0cebf56c3a4158ccd99817eb47b8e9900b1d9e66952dfb9458b154abdb08f85c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LHEw0GBzaW8bUcPXdIPAwG7R/mHeCrrs8STYhMdSBS71PbIhnf+YYSeZaiEKanSAgZOxMyq6oSK9kSfOh2iQDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-13T19:42:05.528323Z","bundle_sha256":"1d95f8f6aa0bca383e226cc89430249396b9944c89d5f33c9e7eb060039014ce"}}