{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QZBU5RTRBOQ7HJ22TLW2X6BPSQ","short_pith_number":"pith:QZBU5RTR","schema_version":"1.0","canonical_sha256":"86434ec6710ba1f3a75a9aedabf82f942a5737198647ce069c28555a9da3fd87","source":{"kind":"arxiv","id":"2406.01468","version":2},"attestation_state":"computed","paper":{"title":"Understanding Token Probability Encoding in Output Embeddings","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Hakaze Cho, Kenshiro Tanaka, Mariko Kato, Naoya Inoue, Yoshihiro Sakai","submitted_at":"2024-06-03T15:57:29Z","abstract_excerpt":"In this paper, we investigate the output token probability information in the output embedding of language models. We find an approximate common log-linear encoding of output token probabilities within the output embedding vectors and empirically demonstrate that it is accurate and sparse. As a causality examination, we steer the encoding in output embedding to modify the output probability distribution accurately. Moreover, the sparsity we find in output probability encoding suggests that a large number of dimensions in the output embedding do not contribute to causal language modeling. There"},"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":"2406.01468","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-03T15:57:29Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"7584d4bd4f25e76550274a434ca4427b04ec3281d67378e819514000f454e028","abstract_canon_sha256":"7ac029e56aab7e080f3a59d411526ca172cf998ec8c7caca704390b81c020efa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:47:30.720501Z","signature_b64":"15cu5XtOFfzpKVvBDpmMfIFbnUeNbLJlS9VlDj8y37Z+W2nmBa5fllT/QGjK0Zd2mW68iOHBrQscBVC1jIJRDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"86434ec6710ba1f3a75a9aedabf82f942a5737198647ce069c28555a9da3fd87","last_reissued_at":"2026-07-05T09:47:30.719861Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:47:30.719861Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding Token Probability Encoding in Output Embeddings","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Hakaze Cho, Kenshiro Tanaka, Mariko Kato, Naoya Inoue, Yoshihiro Sakai","submitted_at":"2024-06-03T15:57:29Z","abstract_excerpt":"In this paper, we investigate the output token probability information in the output embedding of language models. We find an approximate common log-linear encoding of output token probabilities within the output embedding vectors and empirically demonstrate that it is accurate and sparse. As a causality examination, we steer the encoding in output embedding to modify the output probability distribution accurately. Moreover, the sparsity we find in output probability encoding suggests that a large number of dimensions in the output embedding do not contribute to causal language modeling. There"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.01468","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/2406.01468/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":"2406.01468","created_at":"2026-07-05T09:47:30.719939+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.01468v2","created_at":"2026-07-05T09:47:30.719939+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.01468","created_at":"2026-07-05T09:47:30.719939+00:00"},{"alias_kind":"pith_short_12","alias_value":"QZBU5RTRBOQ7","created_at":"2026-07-05T09:47:30.719939+00:00"},{"alias_kind":"pith_short_16","alias_value":"QZBU5RTRBOQ7HJ22","created_at":"2026-07-05T09:47:30.719939+00:00"},{"alias_kind":"pith_short_8","alias_value":"QZBU5RTR","created_at":"2026-07-05T09:47:30.719939+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/QZBU5RTRBOQ7HJ22TLW2X6BPSQ","json":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ.json","graph_json":"https://pith.science/api/pith-number/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/graph.json","events_json":"https://pith.science/api/pith-number/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/events.json","paper":"https://pith.science/paper/QZBU5RTR"},"agent_actions":{"view_html":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ","download_json":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ.json","view_paper":"https://pith.science/paper/QZBU5RTR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.01468&json=true","fetch_graph":"https://pith.science/api/pith-number/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/graph.json","fetch_events":"https://pith.science/api/pith-number/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/action/storage_attestation","attest_author":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/action/author_attestation","sign_citation":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/action/citation_signature","submit_replication":"https://pith.science/pith/QZBU5RTRBOQ7HJ22TLW2X6BPSQ/action/replication_record"}},"created_at":"2026-07-05T09:47:30.719939+00:00","updated_at":"2026-07-05T09:47:30.719939+00:00"}