{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:CKYPY7COKENL7RYMHMYDYQI2JZ","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":"b492531e79708fa65f0e9c1cb0ff89d73220ad2b7dad08d8858f54327095a83e","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T02:55:53Z","title_canon_sha256":"e95d0dd3b25c1ed57ab754e0f7e89c430dd56366b72c4bcbe89f84d95ffec567"},"schema_version":"1.0","source":{"id":"2605.14294","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.14294","created_at":"2026-05-17T23:39:10Z"},{"alias_kind":"arxiv_version","alias_value":"2605.14294v1","created_at":"2026-05-17T23:39:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14294","created_at":"2026-05-17T23:39:10Z"},{"alias_kind":"pith_short_12","alias_value":"CKYPY7COKENL","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"CKYPY7COKENL7RYM","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"CKYPY7CO","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:41860e191ba4e0e1fd845389067d34e1144713a265922bde4d9cba8cef65f671","target":"graph","created_at":"2026-05-17T23:39:10Z","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":"we propose a transformer verification approach that can achieve improved precision... by representing a precise but non-linear bound for dot products such that we can further exploit the rich body of literature for convex relaxation of ReLU to derive precise bounds."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the ReLU encoding of dot-product ranges remains tractable for convex relaxation and that the resulting bounds are tight enough to meaningfully reduce false alarms on the evaluated model sizes and properties."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A ReLU-catalyzed abstraction method yields tighter bounds for transformer verification by converting dot-product constraints into ReLU forms that leverage standard convex relaxations."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"ReLU encoding of dot-product ranges enables tighter convex bounds for precise transformer verification."}],"snapshot_sha256":"01ecd2c9b2580ee639ea2ea8da342a8a05bd0bfa45f8d5340d4f357fe7ddaebe"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Formal verification of transformers has become increasingly important due to their widespread deployment in safety-critical applications. Compared to classic neural networks, the inferences of transformers involve highly complex computations, such as dot products in self-attention layers, rendering their verification extremely difficult. Existing approaches explored over-approximation methods by constructing convex constraints to bound the output ranges of transformers, which can achieve high efficiency. However, they may sacrifice verification precision, and consequently introduce significant","authors_text":"Hengjie Liu, Jianjun Zhao, Zhenya Zhang","cross_cats":["cs.LG"],"headline":"ReLU encoding of dot-product ranges enables tighter convex bounds for precise transformer verification.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T02:55:53Z","title":"Precise Verification of Transformers through ReLU-Catalyzed Abstraction Refinement"},"references":{"count":39,"internal_anchors":2,"resolved_work":39,"sample":[{"cited_arxiv_id":"","doi":"10.18653/v1/d18-1316","is_internal_anchor":false,"ref_index":1,"title":"Alzantot, M., Sharma, Y., Elgohary, A., Ho, B.J., Srivastava, M., Chang, K.W.: Generating natural language adversarial examples. In: Riloff, E., Chiang, D., Hocken- maier, J., Tsujii, J. (eds.) Procee","work_id":"178ae929-ee76-462f-803a-eb9692c82dee","year":2018},{"cited_arxiv_id":"","doi":"10.18653/v1/2021.insights-1.18","is_internal_anchor":false,"ref_index":2,"title":"In: Sedoc, J., Rogers, A., Rumshisky, A., Tafreshi, S","work_id":"1c1c4369-ee00-46d5-99c0-c9512eb150bc","year":2021},{"cited_arxiv_id":"","doi":"10.1145/3453483.3454056","is_internal_anchor":false,"ref_index":3,"title":"Lee, and Brandon Reagen","work_id":"c488edc9-9b8a-4531-a56d-afee25c6b6d3","year":2021},{"cited_arxiv_id":"","doi":"10.1145/3617508","is_internal_anchor":false,"ref_index":4,"title":"Boudardara, F., Boussif, A., Meyer, P.J., Ghazel, M.: A review of abstraction methods toward verifying neural networks. ACM Trans. Embed. Comput. Syst. 23(4) (Jun 2024). https://doi.org/10.1145/361750","work_id":"54f332c1-e698-4bd6-9f3c-96d6dba439b9","year":2024},{"cited_arxiv_id":"2203.14987","doi":"10.48550/arxiv","is_internal_anchor":true,"ref_index":5,"title":"URLhttps://doi.org/10.48550/arXiv","work_id":"5c2060c6-427c-4321-be22-49ccae439d80","year":2024}],"snapshot_sha256":"92da5dace88b78f79caa31ec002320d15769472cbcfc7fb91d06472266fba510"},"source":{"id":"2605.14294","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T02:31:52.684904Z","id":"ef8bad98-1e84-4e9f-b5dc-183f0cb827f3","model_set":{"reader":"grok-4.3"},"one_line_summary":"A ReLU-catalyzed abstraction method yields tighter bounds for transformer verification by converting dot-product constraints into ReLU forms that leverage standard convex relaxations.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"ReLU encoding of dot-product ranges enables tighter convex bounds for precise transformer verification.","strongest_claim":"we propose a transformer verification approach that can achieve improved precision... by representing a precise but non-linear bound for dot products such that we can further exploit the rich body of literature for convex relaxation of ReLU to derive precise bounds.","weakest_assumption":"That the ReLU encoding of dot-product ranges remains tractable for convex relaxation and that the resulting bounds are tight enough to meaningfully reduce false alarms on the evaluated model sizes and properties."}},"verdict_id":"ef8bad98-1e84-4e9f-b5dc-183f0cb827f3"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0076fc4188d26763b79e2d0be8c18d7f39a5aa667bece11c696dd4f3d6dfcd75","target":"record","created_at":"2026-05-17T23:39:10Z","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":"b492531e79708fa65f0e9c1cb0ff89d73220ad2b7dad08d8858f54327095a83e","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T02:55:53Z","title_canon_sha256":"e95d0dd3b25c1ed57ab754e0f7e89c430dd56366b72c4bcbe89f84d95ffec567"},"schema_version":"1.0","source":{"id":"2605.14294","kind":"arxiv","version":1}},"canonical_sha256":"12b0fc7c4e511abfc70c3b303c411a4e4ab8922d463e89f04b20551c2257ba9d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"12b0fc7c4e511abfc70c3b303c411a4e4ab8922d463e89f04b20551c2257ba9d","first_computed_at":"2026-05-17T23:39:10.168048Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:10.168048Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Mapn8k+QFTD0jpfnp6fa7rdg9jxYvvJZkdOBUgvM7kQExVtnzxb8q6RHhqlQpZe/Dj8L2x3tUvIZHLPp0AxMBg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:10.168598Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.14294","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0076fc4188d26763b79e2d0be8c18d7f39a5aa667bece11c696dd4f3d6dfcd75","sha256:41860e191ba4e0e1fd845389067d34e1144713a265922bde4d9cba8cef65f671"],"state_sha256":"340bf5727120dc97e583a19b182b4f42e45b53586c44c67c7020ec3cc13b5a73"}