{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:23ZEPV2YBMQQWAXTUCS4GTVTML","short_pith_number":"pith:23ZEPV2Y","canonical_record":{"source":{"id":"2601.21766","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T14:16:39Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c3fb0b68e932ede11292cdb8cf8460f267c04394665d0887589e98bfb3aef976","abstract_canon_sha256":"cd8be869b46ce6ae8e1dc41fa4026452bb7fb98020eba61eb7978cd0b08053fc"},"schema_version":"1.0"},"canonical_sha256":"d6f247d7580b210b02f3a0a5c34eb362e434871a361617d108dca7ace156604d","source":{"kind":"arxiv","id":"2601.21766","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.21766","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"arxiv_version","alias_value":"2601.21766v4","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.21766","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"pith_short_12","alias_value":"23ZEPV2YBMQQ","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"pith_short_16","alias_value":"23ZEPV2YBMQQWAXT","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"pith_short_8","alias_value":"23ZEPV2Y","created_at":"2026-05-25T02:02:12Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:23ZEPV2YBMQQWAXTUCS4GTVTML","target":"record","payload":{"canonical_record":{"source":{"id":"2601.21766","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T14:16:39Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c3fb0b68e932ede11292cdb8cf8460f267c04394665d0887589e98bfb3aef976","abstract_canon_sha256":"cd8be869b46ce6ae8e1dc41fa4026452bb7fb98020eba61eb7978cd0b08053fc"},"schema_version":"1.0"},"canonical_sha256":"d6f247d7580b210b02f3a0a5c34eb362e434871a361617d108dca7ace156604d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:02:12.893275Z","signature_b64":"gPIKbDDxr9tdY0gyCCsZ7BqpUfnvZQGoBBDrmqzpEnrENbFshrP65/ASBqRaPA9oA18VE3X9THBGUo/6adwdDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d6f247d7580b210b02f3a0a5c34eb362e434871a361617d108dca7ace156604d","last_reissued_at":"2026-05-25T02:02:12.892377Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:02:12.892377Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2601.21766","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-25T02:02:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GXg1BHJJ4YVvSv788q+VNzeXxpz5cJU5OMzKGg/sxBIPQbVuE7Y//ergefp1Nm+WnVKKC5HrU/ATu2t1jUpEDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-08T02:27:59.228052Z"},"content_sha256":"5723a28add6ee4375612c3558f09ef4567ca150c0f956c9ffb86f38f0432993d","schema_version":"1.0","event_id":"sha256:5723a28add6ee4375612c3558f09ef4567ca150c0f956c9ffb86f38f0432993d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:23ZEPV2YBMQQWAXTUCS4GTVTML","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"CoFrGeNet: Continued Fraction Architectures for Language Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Continued-fraction components replace attention and feed-forward layers in large transformers with half to two-thirds the parameters while matching or exceeding performance on language tasks.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Amit Dhurandhar, Dennis Wei, Karthikeyan Natesan Ramamurthy, Rahul Nair, Tejaswini Pedapati, Vijil Chenthamarakshan","submitted_at":"2026-01-29T14:16:39Z","abstract_excerpt":"Transformers are arguably the preferred architecture for language generation. In this paper, inspired by continued fractions, we introduce a new function class for generative modeling. The architecture family implementing this function class is named CoFrGeNets - Continued Fraction Generative Networks. We design novel architectural components based on this function class that can replace Multi-head Attention and Feed-Forward Networks in Transformer blocks while requiring much fewer parameters. We derive custom gradient formulations to optimize the proposed components more accurately and effici"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Results show that the performance on downstream classification, Q&A, reasoning and text understanding tasks of our models is competitive and sometimes even superior to the original models with 2/3 to 1/2 the parameters and shorter pre-training time.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That continued-fraction components can preserve the modeling capacity of attention and feed-forward layers while using far fewer parameters, and that the custom gradient rules produce stable optimization across large-scale pre-training.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"CoFrGeNet uses continued-fraction function classes to build transformer replacements that match or beat GPT-2 and Llama performance with half to two-thirds the parameters.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Continued-fraction components replace attention and feed-forward layers in large transformers with half to two-thirds the parameters while matching or exceeding performance on language tasks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1c791f7da8a642bb60dac7de2216135f80bbd98b8f66719bcea55c9252a80468"},"source":{"id":"2601.21766","kind":"arxiv","version":4},"verdict":{"id":"e3309cf7-b466-427e-a6fb-ccdbd310817e","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T09:43:55.537445Z","strongest_claim":"Results show that the performance on downstream classification, Q&A, reasoning and text understanding tasks of our models is competitive and sometimes even superior to the original models with 2/3 to 1/2 the parameters and shorter pre-training time.","one_line_summary":"CoFrGeNet uses continued-fraction function classes to build transformer replacements that match or beat GPT-2 and Llama performance with half to two-thirds the parameters.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That continued-fraction components can preserve the modeling capacity of attention and feed-forward layers while using far fewer parameters, and that the custom gradient rules produce stable optimization across large-scale pre-training.","pith_extraction_headline":"Continued-fraction components replace attention and feed-forward layers in large transformers with half to two-thirds the parameters while matching or exceeding performance on language tasks."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2601.21766/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":2,"snapshot_sha256":"1420dd19caee143aa61b19d2163d453763b9c16796eb0c9ef10f94c342f5ee1c"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"e3309cf7-b466-427e-a6fb-ccdbd310817e"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-25T02:02:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FJ9NkQN5JEHN/Q4eBjFkiydcLT9EutTsgoB+N8ieQtWFqoUBtxeFg+wshdG7uHnj0Gf/7DnzIT/KMHOVmFDGCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-08T02:27:59.228994Z"},"content_sha256":"22cf66f29da1f5e2dfa713f7b15f75567f723280384a4d33645dabda9f66ca9b","schema_version":"1.0","event_id":"sha256:22cf66f29da1f5e2dfa713f7b15f75567f723280384a4d33645dabda9f66ca9b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/23ZEPV2YBMQQWAXTUCS4GTVTML/bundle.json","state_url":"https://pith.science/pith/23ZEPV2YBMQQWAXTUCS4GTVTML/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/23ZEPV2YBMQQWAXTUCS4GTVTML/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-08T02:27:59Z","links":{"resolver":"https://pith.science/pith/23ZEPV2YBMQQWAXTUCS4GTVTML","bundle":"https://pith.science/pith/23ZEPV2YBMQQWAXTUCS4GTVTML/bundle.json","state":"https://pith.science/pith/23ZEPV2YBMQQWAXTUCS4GTVTML/state.json","well_known_bundle":"https://pith.science/.well-known/pith/23ZEPV2YBMQQWAXTUCS4GTVTML/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:23ZEPV2YBMQQWAXTUCS4GTVTML","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"cd8be869b46ce6ae8e1dc41fa4026452bb7fb98020eba61eb7978cd0b08053fc","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T14:16:39Z","title_canon_sha256":"c3fb0b68e932ede11292cdb8cf8460f267c04394665d0887589e98bfb3aef976"},"schema_version":"1.0","source":{"id":"2601.21766","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.21766","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"arxiv_version","alias_value":"2601.21766v4","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.21766","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"pith_short_12","alias_value":"23ZEPV2YBMQQ","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"pith_short_16","alias_value":"23ZEPV2YBMQQWAXT","created_at":"2026-05-25T02:02:12Z"},{"alias_kind":"pith_short_8","alias_value":"23ZEPV2Y","created_at":"2026-05-25T02:02:12Z"}],"graph_snapshots":[{"event_id":"sha256:22cf66f29da1f5e2dfa713f7b15f75567f723280384a4d33645dabda9f66ca9b","target":"graph","created_at":"2026-05-25T02:02:12Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Results show that the performance on downstream classification, Q&A, reasoning and text understanding tasks of our models is competitive and sometimes even superior to the original models with 2/3 to 1/2 the parameters and shorter pre-training time."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That continued-fraction components can preserve the modeling capacity of attention and feed-forward layers while using far fewer parameters, and that the custom gradient rules produce stable optimization across large-scale pre-training."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"CoFrGeNet uses continued-fraction function classes to build transformer replacements that match or beat GPT-2 and Llama performance with half to two-thirds the parameters."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Continued-fraction components replace attention and feed-forward layers in large transformers with half to two-thirds the parameters while matching or exceeding performance on language tasks."}],"snapshot_sha256":"1c791f7da8a642bb60dac7de2216135f80bbd98b8f66719bcea55c9252a80468"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"1420dd19caee143aa61b19d2163d453763b9c16796eb0c9ef10f94c342f5ee1c"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2601.21766/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Transformers are arguably the preferred architecture for language generation. In this paper, inspired by continued fractions, we introduce a new function class for generative modeling. The architecture family implementing this function class is named CoFrGeNets - Continued Fraction Generative Networks. We design novel architectural components based on this function class that can replace Multi-head Attention and Feed-Forward Networks in Transformer blocks while requiring much fewer parameters. We derive custom gradient formulations to optimize the proposed components more accurately and effici","authors_text":"Amit Dhurandhar, Dennis Wei, Karthikeyan Natesan Ramamurthy, Rahul Nair, Tejaswini Pedapati, Vijil Chenthamarakshan","cross_cats":["cs.AI"],"headline":"Continued-fraction components replace attention and feed-forward layers in large transformers with half to two-thirds the parameters while matching or exceeding performance on language tasks.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T14:16:39Z","title":"CoFrGeNet: Continued Fraction Architectures for Language Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.21766","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-16T09:43:55.537445Z","id":"e3309cf7-b466-427e-a6fb-ccdbd310817e","model_set":{"reader":"grok-4.3"},"one_line_summary":"CoFrGeNet uses continued-fraction function classes to build transformer replacements that match or beat GPT-2 and Llama performance with half to two-thirds the parameters.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Continued-fraction components replace attention and feed-forward layers in large transformers with half to two-thirds the parameters while matching or exceeding performance on language tasks.","strongest_claim":"Results show that the performance on downstream classification, Q&A, reasoning and text understanding tasks of our models is competitive and sometimes even superior to the original models with 2/3 to 1/2 the parameters and shorter pre-training time.","weakest_assumption":"That continued-fraction components can preserve the modeling capacity of attention and feed-forward layers while using far fewer parameters, and that the custom gradient rules produce stable optimization across large-scale pre-training."}},"verdict_id":"e3309cf7-b466-427e-a6fb-ccdbd310817e"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:5723a28add6ee4375612c3558f09ef4567ca150c0f956c9ffb86f38f0432993d","target":"record","created_at":"2026-05-25T02:02:12Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"cd8be869b46ce6ae8e1dc41fa4026452bb7fb98020eba61eb7978cd0b08053fc","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T14:16:39Z","title_canon_sha256":"c3fb0b68e932ede11292cdb8cf8460f267c04394665d0887589e98bfb3aef976"},"schema_version":"1.0","source":{"id":"2601.21766","kind":"arxiv","version":4}},"canonical_sha256":"d6f247d7580b210b02f3a0a5c34eb362e434871a361617d108dca7ace156604d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d6f247d7580b210b02f3a0a5c34eb362e434871a361617d108dca7ace156604d","first_computed_at":"2026-05-25T02:02:12.892377Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-25T02:02:12.892377Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"gPIKbDDxr9tdY0gyCCsZ7BqpUfnvZQGoBBDrmqzpEnrENbFshrP65/ASBqRaPA9oA18VE3X9THBGUo/6adwdDw==","signature_status":"signed_v1","signed_at":"2026-05-25T02:02:12.893275Z","signed_message":"canonical_sha256_bytes"},"source_id":"2601.21766","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5723a28add6ee4375612c3558f09ef4567ca150c0f956c9ffb86f38f0432993d","sha256:22cf66f29da1f5e2dfa713f7b15f75567f723280384a4d33645dabda9f66ca9b"],"state_sha256":"f03072d3a0ef0514491cdee2efc3038f92ff141a0b87f27f219b25dc6e45db95"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XfyReWsY8zZlu0hDXBehJIBl4BOvifsj1axsos9szIUPNp0xFS8UyjyszFE5qEwFuMJdlvrW4shjr9G97yS/AQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-08T02:27:59.233278Z","bundle_sha256":"ff783948b39ef039f608e9b863254300ab2e461a6358982bee9703849173beb7"}}