{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VWY6WYXO3LH6JWA4YBACA4AYEC","short_pith_number":"pith:VWY6WYXO","schema_version":"1.0","canonical_sha256":"adb1eb62eedacfe4d81cc04020701820b6de49ca0b6866e0464d5791fb7ba14b","source":{"kind":"arxiv","id":"2605.25330","version":1},"attestation_state":"computed","paper":{"title":"How Reliable Are Semantic-ID Tokenizer Comparisons in Generative Recommendation?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Haibo Zhang, Jeremiah D. Deng, Lech Szymanski, Qian Zhang","submitted_at":"2026-05-25T01:22:45Z","abstract_excerpt":"In Semantic-ID (SID) based generative recommendation, each item is represented as a sequence of discrete codes, and an autoregressive model is trained to generate the SID sequence of the next item; top-K performance is then measured by checking whether the SID sequence of the target item appears among the generated sequences. This evaluation protocol equates SID-level matching with item-level recommendation, an equivalence that holds only when every SID sequence maps to a single item. We show this assumption breaks down in practice: because tokenizers compress item features into a code space, "},"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":"2605.25330","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2026-05-25T01:22:45Z","cross_cats_sorted":[],"title_canon_sha256":"91336c881b512d8fea032c318aea53a0c2b91e184e5c5711f431dc272cadc056","abstract_canon_sha256":"cdb77ab38a802b729a3b2269f129ebb9dc804f39a979b0e38d164ebb84ea2052"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:04:29.165270Z","signature_b64":"kaEcW61W7m6dBmrQBEZSNdj6hHE+JOAEBPmF1Oru+m9Csp1YOy6eqqRXZzncR9Z7j47vZnc3tnvhqpkN0s3lDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"adb1eb62eedacfe4d81cc04020701820b6de49ca0b6866e0464d5791fb7ba14b","last_reissued_at":"2026-05-26T02:04:29.164600Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:04:29.164600Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Reliable Are Semantic-ID Tokenizer Comparisons in Generative Recommendation?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Haibo Zhang, Jeremiah D. Deng, Lech Szymanski, Qian Zhang","submitted_at":"2026-05-25T01:22:45Z","abstract_excerpt":"In Semantic-ID (SID) based generative recommendation, each item is represented as a sequence of discrete codes, and an autoregressive model is trained to generate the SID sequence of the next item; top-K performance is then measured by checking whether the SID sequence of the target item appears among the generated sequences. This evaluation protocol equates SID-level matching with item-level recommendation, an equivalence that holds only when every SID sequence maps to a single item. We show this assumption breaks down in practice: because tokenizers compress item features into a code space, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25330","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/2605.25330/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":"2605.25330","created_at":"2026-05-26T02:04:29.164719+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.25330v1","created_at":"2026-05-26T02:04:29.164719+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25330","created_at":"2026-05-26T02:04:29.164719+00:00"},{"alias_kind":"pith_short_12","alias_value":"VWY6WYXO3LH6","created_at":"2026-05-26T02:04:29.164719+00:00"},{"alias_kind":"pith_short_16","alias_value":"VWY6WYXO3LH6JWA4","created_at":"2026-05-26T02:04:29.164719+00:00"},{"alias_kind":"pith_short_8","alias_value":"VWY6WYXO","created_at":"2026-05-26T02:04:29.164719+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC","json":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC.json","graph_json":"https://pith.science/api/pith-number/VWY6WYXO3LH6JWA4YBACA4AYEC/graph.json","events_json":"https://pith.science/api/pith-number/VWY6WYXO3LH6JWA4YBACA4AYEC/events.json","paper":"https://pith.science/paper/VWY6WYXO"},"agent_actions":{"view_html":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC","download_json":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC.json","view_paper":"https://pith.science/paper/VWY6WYXO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.25330&json=true","fetch_graph":"https://pith.science/api/pith-number/VWY6WYXO3LH6JWA4YBACA4AYEC/graph.json","fetch_events":"https://pith.science/api/pith-number/VWY6WYXO3LH6JWA4YBACA4AYEC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC/action/storage_attestation","attest_author":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC/action/author_attestation","sign_citation":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC/action/citation_signature","submit_replication":"https://pith.science/pith/VWY6WYXO3LH6JWA4YBACA4AYEC/action/replication_record"}},"created_at":"2026-05-26T02:04:29.164719+00:00","updated_at":"2026-05-26T02:04:29.164719+00:00"}