{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:M5Z7SRPJYOKOD3NLRLOOD4ZPQK","short_pith_number":"pith:M5Z7SRPJ","schema_version":"1.0","canonical_sha256":"6773f945e9c394e1edab8adce1f32f829399c9f2c25841fe06a9718d64c106bb","source":{"kind":"arxiv","id":"2504.17723","version":2},"attestation_state":"computed","paper":{"title":"Statistical Runtime Verification for LLMs via Robustness Estimation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Adiel Ashrov, Guy Katz, Natan Levy","submitted_at":"2025-04-24T16:36:19Z","abstract_excerpt":"Adversarial robustness verification is essential for ensuring the safe deployment of Large Language Models (LLMs) in runtime-critical applications. However, formal verification techniques remain computationally infeasible for modern LLMs due to their exponential runtime and white-box access requirements. This paper presents a case study adapting and extending the RoMA statistical verification framework to assess its feasibility as an online runtime robustness monitor for LLMs in black-box deployment settings. Our adaptation of RoMA analyzes confidence score distributions under semantic perturb"},"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":"2504.17723","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-04-24T16:36:19Z","cross_cats_sorted":[],"title_canon_sha256":"f13733e9d0c6f6dd843fda948fd42c81eb400701ec2a9d6da67a923d35bd2e8c","abstract_canon_sha256":"2edf36c9e27adb6e9f5f0e236f8e67042c660ed75f85f83ff9b0478be93bcca8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:42:28.841869Z","signature_b64":"rly7c4XY+k3nuZdzp+ogtf60NDAGfjB+1aaDpIfQTJdhd50954ETdcRg9maXIKDvXy57uKELgJTyUqZiOL1xBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6773f945e9c394e1edab8adce1f32f829399c9f2c25841fe06a9718d64c106bb","last_reissued_at":"2026-07-05T11:42:28.841356Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:42:28.841356Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Statistical Runtime Verification for LLMs via Robustness Estimation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Adiel Ashrov, Guy Katz, Natan Levy","submitted_at":"2025-04-24T16:36:19Z","abstract_excerpt":"Adversarial robustness verification is essential for ensuring the safe deployment of Large Language Models (LLMs) in runtime-critical applications. However, formal verification techniques remain computationally infeasible for modern LLMs due to their exponential runtime and white-box access requirements. This paper presents a case study adapting and extending the RoMA statistical verification framework to assess its feasibility as an online runtime robustness monitor for LLMs in black-box deployment settings. Our adaptation of RoMA analyzes confidence score distributions under semantic perturb"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.17723","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/2504.17723/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":"2504.17723","created_at":"2026-07-05T11:42:28.841421+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.17723v2","created_at":"2026-07-05T11:42:28.841421+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.17723","created_at":"2026-07-05T11:42:28.841421+00:00"},{"alias_kind":"pith_short_12","alias_value":"M5Z7SRPJYOKO","created_at":"2026-07-05T11:42:28.841421+00:00"},{"alias_kind":"pith_short_16","alias_value":"M5Z7SRPJYOKOD3NL","created_at":"2026-07-05T11:42:28.841421+00:00"},{"alias_kind":"pith_short_8","alias_value":"M5Z7SRPJ","created_at":"2026-07-05T11:42:28.841421+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.27209","citing_title":"Learning to Act under Noise: Enhancing Agent Robustness via Noisy Environments","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09606","citing_title":"Evaluating Reliability Gaps in Large Language Model Safety via Repeated Prompt Sampling","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK","json":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK.json","graph_json":"https://pith.science/api/pith-number/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/graph.json","events_json":"https://pith.science/api/pith-number/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/events.json","paper":"https://pith.science/paper/M5Z7SRPJ"},"agent_actions":{"view_html":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK","download_json":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK.json","view_paper":"https://pith.science/paper/M5Z7SRPJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.17723&json=true","fetch_graph":"https://pith.science/api/pith-number/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/graph.json","fetch_events":"https://pith.science/api/pith-number/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/action/storage_attestation","attest_author":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/action/author_attestation","sign_citation":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/action/citation_signature","submit_replication":"https://pith.science/pith/M5Z7SRPJYOKOD3NLRLOOD4ZPQK/action/replication_record"}},"created_at":"2026-07-05T11:42:28.841421+00:00","updated_at":"2026-07-05T11:42:28.841421+00:00"}