{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:V3UYQNCNSMPWDJ4Y2SFAUDALUE","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":"7caffba25e0399ced70ff7dd358a9b823d9cf9c25485dcf13cf3aace8a5cd998","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-02T02:44:12Z","title_canon_sha256":"aac22fb6e21e389b7878b3d5d35c4bd54f2e77381ceca2f944055dc79d140399"},"schema_version":"1.0","source":{"id":"2605.01205","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.01205","created_at":"2026-06-03T01:05:14Z"},{"alias_kind":"arxiv_version","alias_value":"2605.01205v2","created_at":"2026-06-03T01:05:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.01205","created_at":"2026-06-03T01:05:14Z"},{"alias_kind":"pith_short_12","alias_value":"V3UYQNCNSMPW","created_at":"2026-06-03T01:05:14Z"},{"alias_kind":"pith_short_16","alias_value":"V3UYQNCNSMPWDJ4Y","created_at":"2026-06-03T01:05:14Z"},{"alias_kind":"pith_short_8","alias_value":"V3UYQNCN","created_at":"2026-06-03T01:05:14Z"}],"graph_snapshots":[{"event_id":"sha256:c9b0ec65abb8f5fdfc9eb9cb24923b995987a7e8129314838241240435c7d4c9","target":"graph","created_at":"2026-06-03T01:05:14Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"In challenging cross-architecture distillation experiments, SRA consistently and significantly outperforms state-of-the-art CTKD baselines."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That shifting the alignment unit from tokens to attention-weighted span centers of mass, under a multi-particle dynamical systems framing, produces representations that are both more robust to tokenizer mismatch and more informative for distillation than prior aggregation strategies."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"SRA reframes cross-tokenizer LLM distillation as alignment of attention-weighted span centers of mass in a multi-particle dynamical system and reports consistent gains over prior CTKD baselines."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"SRA shifts LLM distillation alignment from tokens to attention-weighted span centers of mass for better cross-tokenizer transfer."}],"snapshot_sha256":"2b6f216aeeb665cb8f8e001cefaaa7c2377f484267ad41639bb5102aaa89a453"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-20T18:36:46.288022Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T17:29:31.498086Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.01205/integrity.json","findings":[],"snapshot_sha256":"3402b0b2a5fb85a188c8cc23610a5bad512bdc30ddc38005f969a6f1856e0ea7","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Cross-Tokenizer Knowledge Distillation (CTKD) enables knowledge transfer between a large language model and a smaller student, even when they employ different tokenizers. While existing approaches mainly focus on token-level alignment strategies, which are often brittle and sensitive to discrepancies between tokenizers, we argue that the method of aggregating tokens into more robust representations before distillation is of equal importance. In this paper, we introduce \\textbf{SRA} (\\textbf{S}pan \\textbf{R}epresentation \\textbf{A}lignment for Large Language Model Distillation), a novel framewo","authors_text":"Hoang Son Nguyen, Linh Ngo Van, Nguyen Thi Ngoc Diep, Pham Khanh Chi, Quoc Phong Dao, Trung Le, Tung Nguyen","cross_cats":[],"headline":"SRA shifts LLM distillation alignment from tokens to attention-weighted span centers of mass for better cross-tokenizer transfer.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-02T02:44:12Z","title":"SRA: Span Representation Alignment for Large Language Model Distillation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.01205","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-09T15:18:31.469907Z","id":"1f8fbd6d-121d-413c-b7b3-ba799a8d83c9","model_set":{"reader":"grok-4.3"},"one_line_summary":"SRA reframes cross-tokenizer LLM distillation as alignment of attention-weighted span centers of mass in a multi-particle dynamical system and reports consistent gains over prior CTKD baselines.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"SRA shifts LLM distillation alignment from tokens to attention-weighted span centers of mass for better cross-tokenizer transfer.","strongest_claim":"In challenging cross-architecture distillation experiments, SRA consistently and significantly outperforms state-of-the-art CTKD baselines.","weakest_assumption":"That shifting the alignment unit from tokens to attention-weighted span centers of mass, under a multi-particle dynamical systems framing, produces representations that are both more robust to tokenizer mismatch and more informative for distillation than prior aggregation strategies."}},"verdict_id":"1f8fbd6d-121d-413c-b7b3-ba799a8d83c9"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0d93f56e79f65ce6e8ba66feaa90e75b1b7eb296919524fbbda1d3fa6c95162c","target":"record","created_at":"2026-06-03T01:05:14Z","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":"7caffba25e0399ced70ff7dd358a9b823d9cf9c25485dcf13cf3aace8a5cd998","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-02T02:44:12Z","title_canon_sha256":"aac22fb6e21e389b7878b3d5d35c4bd54f2e77381ceca2f944055dc79d140399"},"schema_version":"1.0","source":{"id":"2605.01205","kind":"arxiv","version":2}},"canonical_sha256":"aee988344d931f61a798d48a0a0c0ba12d79aefd38842a37ab602b47029ed2a2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"aee988344d931f61a798d48a0a0c0ba12d79aefd38842a37ab602b47029ed2a2","first_computed_at":"2026-06-03T01:05:14.262900Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-03T01:05:14.262900Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Nvp0NrO0HUj1+4H8Fp9vPR1qmhIi2GsQU85AQrLUVu0hj0H6Z4V1nrV4vqenyYuVlEaoK3/yqMB64lWDtJWKBw==","signature_status":"signed_v1","signed_at":"2026-06-03T01:05:14.263553Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.01205","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0d93f56e79f65ce6e8ba66feaa90e75b1b7eb296919524fbbda1d3fa6c95162c","sha256:c9b0ec65abb8f5fdfc9eb9cb24923b995987a7e8129314838241240435c7d4c9"],"state_sha256":"92420e462db7e211186ba88c7b3ccc3b6863e9799a21a519eda56ecea3904d67"}