{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7XJ36KM2VNBHBAULG6ABTMSR7G","short_pith_number":"pith:7XJ36KM2","schema_version":"1.0","canonical_sha256":"fdd3bf299aab4270828b378019b251f9abe4f47de595c0befe17b275207f5e4a","source":{"kind":"arxiv","id":"2508.08293","version":1},"attestation_state":"computed","paper":{"title":"Topos Theory for Generative AI and LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Sridhar Mahadevan","submitted_at":"2025-08-05T20:00:06Z","abstract_excerpt":"We propose the design of novel categorical generative AI architectures (GAIAs) using topos theory, a type of category that is ``set-like\": a topos has all (co)limits, is Cartesian closed, and has a subobject classifier. Previous theoretical results on the Transformer model have shown that it is a universal sequence-to-sequence function approximator, and dense in the space of all continuous functions with compact support on the Euclidean space of embeddings of tokens. Building on this theoretical result, we explore novel architectures for LLMs that exploit the property that the category of LLMs"},"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":"2508.08293","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-08-05T20:00:06Z","cross_cats_sorted":[],"title_canon_sha256":"b42bbeea3036782f4df5378fa46ea92fb8498d558d066cc54ddc3420540427c2","abstract_canon_sha256":"97975005883149880e8a8b644f1c05b63f5c0deca07547dcaa972c41270cfd34"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:25.417287Z","signature_b64":"jIgSOQmAa3CHPBTomkepm8/hjfCKS9MK7ZrR9EQttKMlWlr8NQv1qa+2ZKzVTyd7EXhIh0CoQwvo7LULKyAlAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fdd3bf299aab4270828b378019b251f9abe4f47de595c0befe17b275207f5e4a","last_reissued_at":"2026-07-05T11:52:25.416808Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:25.416808Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Topos Theory for Generative AI and LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Sridhar Mahadevan","submitted_at":"2025-08-05T20:00:06Z","abstract_excerpt":"We propose the design of novel categorical generative AI architectures (GAIAs) using topos theory, a type of category that is ``set-like\": a topos has all (co)limits, is Cartesian closed, and has a subobject classifier. Previous theoretical results on the Transformer model have shown that it is a universal sequence-to-sequence function approximator, and dense in the space of all continuous functions with compact support on the Euclidean space of embeddings of tokens. Building on this theoretical result, we explore novel architectures for LLMs that exploit the property that the category of LLMs"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.08293","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/2508.08293/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":"2508.08293","created_at":"2026-07-05T11:52:25.416868+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.08293v1","created_at":"2026-07-05T11:52:25.416868+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.08293","created_at":"2026-07-05T11:52:25.416868+00:00"},{"alias_kind":"pith_short_12","alias_value":"7XJ36KM2VNBH","created_at":"2026-07-05T11:52:25.416868+00:00"},{"alias_kind":"pith_short_16","alias_value":"7XJ36KM2VNBHBAUL","created_at":"2026-07-05T11:52:25.416868+00:00"},{"alias_kind":"pith_short_8","alias_value":"7XJ36KM2","created_at":"2026-07-05T11:52:25.416868+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.28864","citing_title":"The Cognitive Categorical Transformer: Category-Theoretic Inductive Biases for Language Modeling","ref_index":19,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G","json":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G.json","graph_json":"https://pith.science/api/pith-number/7XJ36KM2VNBHBAULG6ABTMSR7G/graph.json","events_json":"https://pith.science/api/pith-number/7XJ36KM2VNBHBAULG6ABTMSR7G/events.json","paper":"https://pith.science/paper/7XJ36KM2"},"agent_actions":{"view_html":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G","download_json":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G.json","view_paper":"https://pith.science/paper/7XJ36KM2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.08293&json=true","fetch_graph":"https://pith.science/api/pith-number/7XJ36KM2VNBHBAULG6ABTMSR7G/graph.json","fetch_events":"https://pith.science/api/pith-number/7XJ36KM2VNBHBAULG6ABTMSR7G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G/action/storage_attestation","attest_author":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G/action/author_attestation","sign_citation":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G/action/citation_signature","submit_replication":"https://pith.science/pith/7XJ36KM2VNBHBAULG6ABTMSR7G/action/replication_record"}},"created_at":"2026-07-05T11:52:25.416868+00:00","updated_at":"2026-07-05T11:52:25.416868+00:00"}