{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SFR52VHZXNODG3SGEQJDIDPQ2C","short_pith_number":"pith:SFR52VHZ","schema_version":"1.0","canonical_sha256":"9163dd54f9bb5c336e462412340df0d0b4091eb194d25d79aeef4ad5df276167","source":{"kind":"arxiv","id":"2409.12924","version":4},"attestation_state":"computed","paper":{"title":"Wavelet GPT: Wavelet Inspired Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG","cs.SD","eess.AS"],"primary_cat":"eess.SP","authors_text":"Prateek Verma","submitted_at":"2024-09-04T03:17:19Z","abstract_excerpt":"Large Language Models (LLMs) have ushered in a new wave of artificial intelligence advancements impacting every scientific field and discipline. We live in a world where most of the data around us, e.g., text, audio, and music, has a multi-scale structure. This paper infuses LLMs with a traditional signal processing idea, namely wavelets, during pre-training to take advantage of the structure. Without adding \\textbf{any extra parameters} to a GPT-style LLM architecture in an academic setup, we achieve the same pre-training performance almost twice as fast in text, audio, and images. This is do"},"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":"2409.12924","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.SP","submitted_at":"2024-09-04T03:17:19Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG","cs.SD","eess.AS"],"title_canon_sha256":"34047c43c1041d1d9405f6d84428b05db9210984cdc41256a49ad7e782a0a531","abstract_canon_sha256":"0750a3df8bbd3a50b5c1c9dca49feb177f3635d4f3c74d0bdfdd5c6b072253e2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:11:24.447026Z","signature_b64":"fiJBIR1YV/XlykSQNA0iWEXKUUWFhPF2n4f5VtLiQUOfx8/qJm8kZLkiQLA6OobP/uTdH4oZqsfH6c+CLBadCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9163dd54f9bb5c336e462412340df0d0b4091eb194d25d79aeef4ad5df276167","last_reissued_at":"2026-07-05T10:11:24.446452Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:11:24.446452Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Wavelet GPT: Wavelet Inspired Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG","cs.SD","eess.AS"],"primary_cat":"eess.SP","authors_text":"Prateek Verma","submitted_at":"2024-09-04T03:17:19Z","abstract_excerpt":"Large Language Models (LLMs) have ushered in a new wave of artificial intelligence advancements impacting every scientific field and discipline. We live in a world where most of the data around us, e.g., text, audio, and music, has a multi-scale structure. This paper infuses LLMs with a traditional signal processing idea, namely wavelets, during pre-training to take advantage of the structure. Without adding \\textbf{any extra parameters} to a GPT-style LLM architecture in an academic setup, we achieve the same pre-training performance almost twice as fast in text, audio, and images. This is do"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.12924","kind":"arxiv","version":4},"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/2409.12924/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":"2409.12924","created_at":"2026-07-05T10:11:24.446512+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.12924v4","created_at":"2026-07-05T10:11:24.446512+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.12924","created_at":"2026-07-05T10:11:24.446512+00:00"},{"alias_kind":"pith_short_12","alias_value":"SFR52VHZXNOD","created_at":"2026-07-05T10:11:24.446512+00:00"},{"alias_kind":"pith_short_16","alias_value":"SFR52VHZXNODG3SG","created_at":"2026-07-05T10:11:24.446512+00:00"},{"alias_kind":"pith_short_8","alias_value":"SFR52VHZ","created_at":"2026-07-05T10:11:24.446512+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07478","citing_title":"FourierQK: Spectral Preprocessing of Query-Key Projections Improves Transformer Attention","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C","json":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C.json","graph_json":"https://pith.science/api/pith-number/SFR52VHZXNODG3SGEQJDIDPQ2C/graph.json","events_json":"https://pith.science/api/pith-number/SFR52VHZXNODG3SGEQJDIDPQ2C/events.json","paper":"https://pith.science/paper/SFR52VHZ"},"agent_actions":{"view_html":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C","download_json":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C.json","view_paper":"https://pith.science/paper/SFR52VHZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.12924&json=true","fetch_graph":"https://pith.science/api/pith-number/SFR52VHZXNODG3SGEQJDIDPQ2C/graph.json","fetch_events":"https://pith.science/api/pith-number/SFR52VHZXNODG3SGEQJDIDPQ2C/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C/action/storage_attestation","attest_author":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C/action/author_attestation","sign_citation":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C/action/citation_signature","submit_replication":"https://pith.science/pith/SFR52VHZXNODG3SGEQJDIDPQ2C/action/replication_record"}},"created_at":"2026-07-05T10:11:24.446512+00:00","updated_at":"2026-07-05T10:11:24.446512+00:00"}