{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:EOHBJ7MRTCCLPUUO7DS7Y4ZUOB","short_pith_number":"pith:EOHBJ7MR","schema_version":"1.0","canonical_sha256":"238e14fd919884b7d28ef8e5fc7334704bcaf5416dae7f347e619b955f85e459","source":{"kind":"arxiv","id":"2501.00659","version":2},"attestation_state":"computed","paper":{"title":"Why Are Positional Encodings Nonessential for Deep Autoregressive Transformers? Revisiting a Petroglyph","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Kazuki Irie","submitted_at":"2024-12-31T22:12:45Z","abstract_excerpt":"Do autoregressive Transformer language models require explicit positional encodings (PEs)? The answer is 'no' provided they have more than one layer -- they can distinguish sequences with permuted tokens without the need for explicit PEs. This follows from the fact that a cascade of (permutation invariant) set processors can collectively exhibit sequence-sensitive behavior in the autoregressive setting. This property has been known since early efforts (contemporary with GPT-2) adopting the Transformer for language modeling. However, this result does not appear to have been well disseminated, l"},"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":"2501.00659","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-12-31T22:12:45Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"c8e10de760b47b81e7d9a55c21cbfcb2b3372f523b9aece81abcd56eafd634fe","abstract_canon_sha256":"5b40f31959c6f08d188e458e3869964bd8fe4b45fe330dd3f54b3625511156bf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:17.923681Z","signature_b64":"VuvpGyFJjFZfhc/rvV/18E5kn3KWEXQvGHtbWmc3MKnMMejBV5kXsArKHJBLmvkqQY/G/U/QldXkCJbU3zF8CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"238e14fd919884b7d28ef8e5fc7334704bcaf5416dae7f347e619b955f85e459","last_reissued_at":"2026-07-05T11:13:17.923225Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:17.923225Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Why Are Positional Encodings Nonessential for Deep Autoregressive Transformers? Revisiting a Petroglyph","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Kazuki Irie","submitted_at":"2024-12-31T22:12:45Z","abstract_excerpt":"Do autoregressive Transformer language models require explicit positional encodings (PEs)? The answer is 'no' provided they have more than one layer -- they can distinguish sequences with permuted tokens without the need for explicit PEs. This follows from the fact that a cascade of (permutation invariant) set processors can collectively exhibit sequence-sensitive behavior in the autoregressive setting. This property has been known since early efforts (contemporary with GPT-2) adopting the Transformer for language modeling. However, this result does not appear to have been well disseminated, l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.00659","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/2501.00659/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":"2501.00659","created_at":"2026-07-05T11:13:17.923280+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.00659v2","created_at":"2026-07-05T11:13:17.923280+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.00659","created_at":"2026-07-05T11:13:17.923280+00:00"},{"alias_kind":"pith_short_12","alias_value":"EOHBJ7MRTCCL","created_at":"2026-07-05T11:13:17.923280+00:00"},{"alias_kind":"pith_short_16","alias_value":"EOHBJ7MRTCCLPUUO","created_at":"2026-07-05T11:13:17.923280+00:00"},{"alias_kind":"pith_short_8","alias_value":"EOHBJ7MR","created_at":"2026-07-05T11:13:17.923280+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2511.19693","citing_title":"TREASURE: The Visa Payment Foundation Model for High-Volume Transaction Understanding","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB","json":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB.json","graph_json":"https://pith.science/api/pith-number/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/graph.json","events_json":"https://pith.science/api/pith-number/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/events.json","paper":"https://pith.science/paper/EOHBJ7MR"},"agent_actions":{"view_html":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB","download_json":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB.json","view_paper":"https://pith.science/paper/EOHBJ7MR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.00659&json=true","fetch_graph":"https://pith.science/api/pith-number/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/graph.json","fetch_events":"https://pith.science/api/pith-number/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/action/storage_attestation","attest_author":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/action/author_attestation","sign_citation":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/action/citation_signature","submit_replication":"https://pith.science/pith/EOHBJ7MRTCCLPUUO7DS7Y4ZUOB/action/replication_record"}},"created_at":"2026-07-05T11:13:17.923280+00:00","updated_at":"2026-07-05T11:13:17.923280+00:00"}