{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:XLC5NG7ZMVHIL7M4HUFZAFUGPV","short_pith_number":"pith:XLC5NG7Z","canonical_record":{"source":{"id":"2403.09635","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-03-14T17:59:14Z","cross_cats_sorted":["cs.AI","cs.CV","cs.LG"],"title_canon_sha256":"fda40688546dba339ca0fdc49cf682bb6d1b919cedc8dac93fb3215a69aacb51","abstract_canon_sha256":"992db4f775f3786630ce44ffe911a683b938cfff5a048a89e0547d68c0dcf50f"},"schema_version":"1.0"},"canonical_sha256":"bac5d69bf9654e85fd9c3d0b9016867d478a26ceb11f37d07f228104252855a6","source":{"kind":"arxiv","id":"2403.09635","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.09635","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"arxiv_version","alias_value":"2403.09635v2","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.09635","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"pith_short_12","alias_value":"XLC5NG7ZMVHI","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"pith_short_16","alias_value":"XLC5NG7ZMVHIL7M4","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"pith_short_8","alias_value":"XLC5NG7Z","created_at":"2026-07-05T08:45:26Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:XLC5NG7ZMVHIL7M4HUFZAFUGPV","target":"record","payload":{"canonical_record":{"source":{"id":"2403.09635","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-03-14T17:59:14Z","cross_cats_sorted":["cs.AI","cs.CV","cs.LG"],"title_canon_sha256":"fda40688546dba339ca0fdc49cf682bb6d1b919cedc8dac93fb3215a69aacb51","abstract_canon_sha256":"992db4f775f3786630ce44ffe911a683b938cfff5a048a89e0547d68c0dcf50f"},"schema_version":"1.0"},"canonical_sha256":"bac5d69bf9654e85fd9c3d0b9016867d478a26ceb11f37d07f228104252855a6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:45:26.312544Z","signature_b64":"/IqPh1oyQCl6UhCBzdoOBu5Zqd4c2aqsh7EaySKWJU2DXX8HeyHKbx6LRmY41dTiYP8VkwSw7hm5voDBvFH6AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bac5d69bf9654e85fd9c3d0b9016867d478a26ceb11f37d07f228104252855a6","last_reissued_at":"2026-07-05T08:45:26.312007Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:45:26.312007Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2403.09635","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:45:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7zTxBxP8Udr9GqYyfado/1Z8t8eejKJBSrsSzXUzBDIxYlm0ZyCCNq6W5vOT3887FErPWVayMUimi9Q5fAGwBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T08:14:21.004415Z"},"content_sha256":"8aeae8b3aedbe4ee575d9406d994778f69d20e83354490aec11fff7ddc1a97f9","schema_version":"1.0","event_id":"sha256:8aeae8b3aedbe4ee575d9406d994778f69d20e83354490aec11fff7ddc1a97f9"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:XLC5NG7ZMVHIL7M4HUFZAFUGPV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Transformers Get Stable: An End-to-End Signal Propagation Theory for Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Akhil Kedia, Haejun Lee, Harshith Goka, Jungho Jung, Mohd Abbas Zaidi, Sushil Khyalia","submitted_at":"2024-03-14T17:59:14Z","abstract_excerpt":"In spite of their huge success, transformer models remain difficult to scale in depth. In this work, we develop a unified signal propagation theory and provide formulae that govern the moments of the forward and backward signal through the transformer model. Our framework can be used to understand and mitigate vanishing/exploding gradients, rank collapse, and instability associated with high attention scores. We also propose DeepScaleLM, an initialization and scaling scheme that conserves unit output/gradient moments throughout the model, enabling the training of very deep models with 1000 lay"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.09635","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/2403.09635/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:45:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vJVBe6kIyA2r94fb0DX8xGh4UZYQX61TB9wOB8WIVQ5kYQeWVXLVgEswKr9OFWs5K0m7FE7cBxNXZq2UgYsZDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T08:14:21.004917Z"},"content_sha256":"213a2138979f70d069c24c3f79e3600a8f69ae3fe7c8f31cfe6d10cafa16bd4c","schema_version":"1.0","event_id":"sha256:213a2138979f70d069c24c3f79e3600a8f69ae3fe7c8f31cfe6d10cafa16bd4c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/XLC5NG7ZMVHIL7M4HUFZAFUGPV/bundle.json","state_url":"https://pith.science/pith/XLC5NG7ZMVHIL7M4HUFZAFUGPV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/XLC5NG7ZMVHIL7M4HUFZAFUGPV/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":"2026-08-04T08:14:21Z","links":{"resolver":"https://pith.science/pith/XLC5NG7ZMVHIL7M4HUFZAFUGPV","bundle":"https://pith.science/pith/XLC5NG7ZMVHIL7M4HUFZAFUGPV/bundle.json","state":"https://pith.science/pith/XLC5NG7ZMVHIL7M4HUFZAFUGPV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/XLC5NG7ZMVHIL7M4HUFZAFUGPV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:XLC5NG7ZMVHIL7M4HUFZAFUGPV","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":"992db4f775f3786630ce44ffe911a683b938cfff5a048a89e0547d68c0dcf50f","cross_cats_sorted":["cs.AI","cs.CV","cs.LG"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-03-14T17:59:14Z","title_canon_sha256":"fda40688546dba339ca0fdc49cf682bb6d1b919cedc8dac93fb3215a69aacb51"},"schema_version":"1.0","source":{"id":"2403.09635","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.09635","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"arxiv_version","alias_value":"2403.09635v2","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.09635","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"pith_short_12","alias_value":"XLC5NG7ZMVHI","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"pith_short_16","alias_value":"XLC5NG7ZMVHIL7M4","created_at":"2026-07-05T08:45:26Z"},{"alias_kind":"pith_short_8","alias_value":"XLC5NG7Z","created_at":"2026-07-05T08:45:26Z"}],"graph_snapshots":[{"event_id":"sha256:213a2138979f70d069c24c3f79e3600a8f69ae3fe7c8f31cfe6d10cafa16bd4c","target":"graph","created_at":"2026-07-05T08:45:26Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2403.09635/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In spite of their huge success, transformer models remain difficult to scale in depth. In this work, we develop a unified signal propagation theory and provide formulae that govern the moments of the forward and backward signal through the transformer model. Our framework can be used to understand and mitigate vanishing/exploding gradients, rank collapse, and instability associated with high attention scores. We also propose DeepScaleLM, an initialization and scaling scheme that conserves unit output/gradient moments throughout the model, enabling the training of very deep models with 1000 lay","authors_text":"Akhil Kedia, Haejun Lee, Harshith Goka, Jungho Jung, Mohd Abbas Zaidi, Sushil Khyalia","cross_cats":["cs.AI","cs.CV","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-03-14T17:59:14Z","title":"Transformers Get Stable: An End-to-End Signal Propagation Theory for Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.09635","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:8aeae8b3aedbe4ee575d9406d994778f69d20e83354490aec11fff7ddc1a97f9","target":"record","created_at":"2026-07-05T08:45:26Z","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":"992db4f775f3786630ce44ffe911a683b938cfff5a048a89e0547d68c0dcf50f","cross_cats_sorted":["cs.AI","cs.CV","cs.LG"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-03-14T17:59:14Z","title_canon_sha256":"fda40688546dba339ca0fdc49cf682bb6d1b919cedc8dac93fb3215a69aacb51"},"schema_version":"1.0","source":{"id":"2403.09635","kind":"arxiv","version":2}},"canonical_sha256":"bac5d69bf9654e85fd9c3d0b9016867d478a26ceb11f37d07f228104252855a6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"bac5d69bf9654e85fd9c3d0b9016867d478a26ceb11f37d07f228104252855a6","first_computed_at":"2026-07-05T08:45:26.312007Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:45:26.312007Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/IqPh1oyQCl6UhCBzdoOBu5Zqd4c2aqsh7EaySKWJU2DXX8HeyHKbx6LRmY41dTiYP8VkwSw7hm5voDBvFH6AA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:45:26.312544Z","signed_message":"canonical_sha256_bytes"},"source_id":"2403.09635","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8aeae8b3aedbe4ee575d9406d994778f69d20e83354490aec11fff7ddc1a97f9","sha256:213a2138979f70d069c24c3f79e3600a8f69ae3fe7c8f31cfe6d10cafa16bd4c"],"state_sha256":"385a9999e5d9ade444bff3fc2b9f0b90ee1ad6e7d042e6dc4ab72046b5142ec0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MuJ9QjHr85/jvZsh0q/y2SamM2cjjNsYRylVsCYDf72wlRqbCxt+D1sDlpx7DHTYG3PntG0DQ10MVMhi4IJYCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T08:14:21.008442Z","bundle_sha256":"9c899c7a6e430b63a9756c34d20ef661fb434231cee422cbc6475c806ad2e492"}}