{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:JMZQR3UAQKTF2BQ4P2DXSLCFWL","short_pith_number":"pith:JMZQR3UA","canonical_record":{"source":{"id":"2302.06461","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-13T15:41:20Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"261657ef833ee136ae75961c03d198554c6655a577d82d1e62aa5027ed84434f","abstract_canon_sha256":"75c263de82c96dc6e881253731155c4509a7e18fdc38b45f554b106adc7e5f40"},"schema_version":"1.0"},"canonical_sha256":"4b3308ee8082a65d061c7e87792c45b2c9d3f973c1e95769aec53e4a1c9936db","source":{"kind":"arxiv","id":"2302.06461","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2302.06461","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"arxiv_version","alias_value":"2302.06461v1","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.06461","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"pith_short_12","alias_value":"JMZQR3UAQKTF","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"pith_short_16","alias_value":"JMZQR3UAQKTF2BQ4","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"pith_short_8","alias_value":"JMZQR3UA","created_at":"2026-07-05T05:41:05Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:JMZQR3UAQKTF2BQ4P2DXSLCFWL","target":"record","payload":{"canonical_record":{"source":{"id":"2302.06461","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-13T15:41:20Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"261657ef833ee136ae75961c03d198554c6655a577d82d1e62aa5027ed84434f","abstract_canon_sha256":"75c263de82c96dc6e881253731155c4509a7e18fdc38b45f554b106adc7e5f40"},"schema_version":"1.0"},"canonical_sha256":"4b3308ee8082a65d061c7e87792c45b2c9d3f973c1e95769aec53e4a1c9936db","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:41:05.503321Z","signature_b64":"9XoQ0DJBbhWweg1EbweNDDijgOd5PsqYl1SzOZOPfFbucYObejNPMnv5AdU3yA0vRgnnaGS4fBXeQwGFgQ5VCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4b3308ee8082a65d061c7e87792c45b2c9d3f973c1e95769aec53e4a1c9936db","last_reissued_at":"2026-07-05T05:41:05.502883Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:41:05.502883Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2302.06461","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:41:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HrqDs9Gi7Xe7HsHkk9faELT9qG0E7xGiJ+ipQf8AXsWtNWhLVCKFXx2rXqnUYDPuJdg329S5yCGZZEgfJNksDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T07:11:00.473542Z"},"content_sha256":"c23f781e2a76300e5fff05713ebea86a8fdc7cfe2cfcf86cfb57907f636ec1e4","schema_version":"1.0","event_id":"sha256:c23f781e2a76300e5fff05713ebea86a8fdc7cfe2cfcf86cfb57907f636ec1e4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:JMZQR3UAQKTF2BQ4P2DXSLCFWL","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"A Study on ReLU and Softmax in Transformer","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Jiang Bian, Junliang Guo, Kai Shen, Rui Wang, Siliang Tang, Xu Tan","submitted_at":"2023-02-13T15:41:20Z","abstract_excerpt":"The Transformer architecture consists of self-attention and feed-forward networks (FFNs) which can be viewed as key-value memories according to previous works. However, FFN and traditional memory utilize different activation functions (i.e., ReLU and Softmax respectively), which makes them not equivalent. In this paper, we first rebuild the connections between FFN and key-value memory by conducting extensive studies on ReLU and Softmax, and find they are equivalent when adding an additional layer normalization module on Softmax. In addition, ReLU outperforms Softmax on both FFN and key-value m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.06461","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/2302.06461/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":"2026-07-05T05:41:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FhPAjbiK7J0ZxgqplH3B/VEi6i8PnX80Nk5svO6JN2rk8Dx0DM91ylSzT+jQb5e2mdalIXwDwSvxSUSklUrSCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T07:11:00.474068Z"},"content_sha256":"4e562b80578a8ef717c5a06320854276fbd54f7dd804de0b371f64aa6aa9ddde","schema_version":"1.0","event_id":"sha256:4e562b80578a8ef717c5a06320854276fbd54f7dd804de0b371f64aa6aa9ddde"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JMZQR3UAQKTF2BQ4P2DXSLCFWL/bundle.json","state_url":"https://pith.science/pith/JMZQR3UAQKTF2BQ4P2DXSLCFWL/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JMZQR3UAQKTF2BQ4P2DXSLCFWL/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":"2026-08-05T07:11:00Z","links":{"resolver":"https://pith.science/pith/JMZQR3UAQKTF2BQ4P2DXSLCFWL","bundle":"https://pith.science/pith/JMZQR3UAQKTF2BQ4P2DXSLCFWL/bundle.json","state":"https://pith.science/pith/JMZQR3UAQKTF2BQ4P2DXSLCFWL/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JMZQR3UAQKTF2BQ4P2DXSLCFWL/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:JMZQR3UAQKTF2BQ4P2DXSLCFWL","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":"75c263de82c96dc6e881253731155c4509a7e18fdc38b45f554b106adc7e5f40","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-13T15:41:20Z","title_canon_sha256":"261657ef833ee136ae75961c03d198554c6655a577d82d1e62aa5027ed84434f"},"schema_version":"1.0","source":{"id":"2302.06461","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2302.06461","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"arxiv_version","alias_value":"2302.06461v1","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.06461","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"pith_short_12","alias_value":"JMZQR3UAQKTF","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"pith_short_16","alias_value":"JMZQR3UAQKTF2BQ4","created_at":"2026-07-05T05:41:05Z"},{"alias_kind":"pith_short_8","alias_value":"JMZQR3UA","created_at":"2026-07-05T05:41:05Z"}],"graph_snapshots":[{"event_id":"sha256:4e562b80578a8ef717c5a06320854276fbd54f7dd804de0b371f64aa6aa9ddde","target":"graph","created_at":"2026-07-05T05:41:05Z","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/2302.06461/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The Transformer architecture consists of self-attention and feed-forward networks (FFNs) which can be viewed as key-value memories according to previous works. However, FFN and traditional memory utilize different activation functions (i.e., ReLU and Softmax respectively), which makes them not equivalent. In this paper, we first rebuild the connections between FFN and key-value memory by conducting extensive studies on ReLU and Softmax, and find they are equivalent when adding an additional layer normalization module on Softmax. In addition, ReLU outperforms Softmax on both FFN and key-value m","authors_text":"Jiang Bian, Junliang Guo, Kai Shen, Rui Wang, Siliang Tang, Xu Tan","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-13T15:41:20Z","title":"A Study on ReLU and Softmax in Transformer"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.06461","kind":"arxiv","version":1},"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:c23f781e2a76300e5fff05713ebea86a8fdc7cfe2cfcf86cfb57907f636ec1e4","target":"record","created_at":"2026-07-05T05:41:05Z","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":"75c263de82c96dc6e881253731155c4509a7e18fdc38b45f554b106adc7e5f40","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-13T15:41:20Z","title_canon_sha256":"261657ef833ee136ae75961c03d198554c6655a577d82d1e62aa5027ed84434f"},"schema_version":"1.0","source":{"id":"2302.06461","kind":"arxiv","version":1}},"canonical_sha256":"4b3308ee8082a65d061c7e87792c45b2c9d3f973c1e95769aec53e4a1c9936db","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4b3308ee8082a65d061c7e87792c45b2c9d3f973c1e95769aec53e4a1c9936db","first_computed_at":"2026-07-05T05:41:05.502883Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:41:05.502883Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"9XoQ0DJBbhWweg1EbweNDDijgOd5PsqYl1SzOZOPfFbucYObejNPMnv5AdU3yA0vRgnnaGS4fBXeQwGFgQ5VCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T05:41:05.503321Z","signed_message":"canonical_sha256_bytes"},"source_id":"2302.06461","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c23f781e2a76300e5fff05713ebea86a8fdc7cfe2cfcf86cfb57907f636ec1e4","sha256:4e562b80578a8ef717c5a06320854276fbd54f7dd804de0b371f64aa6aa9ddde"],"state_sha256":"8ed0531a74e92fb15cae71816863ba3faf7f2c0c32b17ee50fb820a32de2d449"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Xt80byLNgy7c4qQaXWUIwwDYKTPBNstmEMEr2irmKPvZjGCM7jDdM4+3Pjb6R8qOfxYk/VVG5mQPmiiPaTyiDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T07:11:00.479211Z","bundle_sha256":"8bfbe00fb8bebd44d81f30b5e301d13ade2121a3c94cb7a94f937eb4dd7d5d8d"}}