{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ANW7IZK3EWF3BT4IULWKCKDIX4","short_pith_number":"pith:ANW7IZK3","schema_version":"1.0","canonical_sha256":"036df4655b258bb0cf88a2eca12868bf1847c1b46e84483587f9e1790ac147fe","source":{"kind":"arxiv","id":"2509.23806","version":2},"attestation_state":"computed","paper":{"title":"Influence-Guided Concolic Testing of Transformer Robustness","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.SE","authors_text":"Chih-Cheng Yang, Chih-Duo Hong, Fang Yu, Yu Wang","submitted_at":"2025-09-28T11:09:15Z","abstract_excerpt":"Concolic testing for neural networks alternates concrete execution with constraint solving to search for inputs that flip model decisions. We present a concolic tester for Transformer classifiers that uses SHAP estimates to rank pending path predicates by their impact on the current prediction. To support self-attention with multiple heads in execution backed by SMT solving, we implement attention semantics in pure Python that are compatible with the solver and make the softmax boundary explicit by concretizing exponentiation arguments. We evaluate our method on CIFAR-10 across three compact T"},"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":"2509.23806","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2025-09-28T11:09:15Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"754f9b89311cc8ead63fa0700fc34e6b861bd6b920e6354693d337c9b0840981","abstract_canon_sha256":"06614e199b71e7957b79bf30470d62a7d3d3a8c40f3fa6891df504fc02b6ca1f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:12:15.151391Z","signature_b64":"YMUwpdZi5unnjvl6TKBVW+cgP3PeVvBl1ekFvNzrpeKi2hbCpx8XMw4Pnu/I0VaRSzm2z+gBaAUpcCnCTodwCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"036df4655b258bb0cf88a2eca12868bf1847c1b46e84483587f9e1790ac147fe","last_reissued_at":"2026-06-19T16:12:15.151036Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:12:15.151036Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Influence-Guided Concolic Testing of Transformer Robustness","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.SE","authors_text":"Chih-Cheng Yang, Chih-Duo Hong, Fang Yu, Yu Wang","submitted_at":"2025-09-28T11:09:15Z","abstract_excerpt":"Concolic testing for neural networks alternates concrete execution with constraint solving to search for inputs that flip model decisions. We present a concolic tester for Transformer classifiers that uses SHAP estimates to rank pending path predicates by their impact on the current prediction. To support self-attention with multiple heads in execution backed by SMT solving, we implement attention semantics in pure Python that are compatible with the solver and make the softmax boundary explicit by concretizing exponentiation arguments. We evaluate our method on CIFAR-10 across three compact T"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.23806","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/2509.23806/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":"2509.23806","created_at":"2026-06-19T16:12:15.151092+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.23806v2","created_at":"2026-06-19T16:12:15.151092+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.23806","created_at":"2026-06-19T16:12:15.151092+00:00"},{"alias_kind":"pith_short_12","alias_value":"ANW7IZK3EWF3","created_at":"2026-06-19T16:12:15.151092+00:00"},{"alias_kind":"pith_short_16","alias_value":"ANW7IZK3EWF3BT4I","created_at":"2026-06-19T16:12:15.151092+00:00"},{"alias_kind":"pith_short_8","alias_value":"ANW7IZK3","created_at":"2026-06-19T16:12:15.151092+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.26545","citing_title":"ConcoLixir: Reactive LLM Discovery Oracles for Python Concolic Testing","ref_index":15,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4","json":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4.json","graph_json":"https://pith.science/api/pith-number/ANW7IZK3EWF3BT4IULWKCKDIX4/graph.json","events_json":"https://pith.science/api/pith-number/ANW7IZK3EWF3BT4IULWKCKDIX4/events.json","paper":"https://pith.science/paper/ANW7IZK3"},"agent_actions":{"view_html":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4","download_json":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4.json","view_paper":"https://pith.science/paper/ANW7IZK3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.23806&json=true","fetch_graph":"https://pith.science/api/pith-number/ANW7IZK3EWF3BT4IULWKCKDIX4/graph.json","fetch_events":"https://pith.science/api/pith-number/ANW7IZK3EWF3BT4IULWKCKDIX4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4/action/storage_attestation","attest_author":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4/action/author_attestation","sign_citation":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4/action/citation_signature","submit_replication":"https://pith.science/pith/ANW7IZK3EWF3BT4IULWKCKDIX4/action/replication_record"}},"created_at":"2026-06-19T16:12:15.151092+00:00","updated_at":"2026-06-19T16:12:15.151092+00:00"}