{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:WBKMFDVLOIVBUGMRVZPE7OLTWZ","short_pith_number":"pith:WBKMFDVL","schema_version":"1.0","canonical_sha256":"b054c28eab722a1a1991ae5e4fb973b65f2e5decd6610cbe74df8428686aa577","source":{"kind":"arxiv","id":"2504.08798","version":1},"attestation_state":"computed","paper":{"title":"Exploring Gradient-Guided Masked Language Model to Detect Textual Adversarial Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Leo Yu Zhang, Shengshan Hu, Shirui Pan, Xiaomei Zhang, Xufei Zheng, Yanjun Zhang, Zhaoxi Zhang","submitted_at":"2025-04-08T14:10:57Z","abstract_excerpt":"Textual adversarial examples pose serious threats to the reliability of natural language processing systems. Recent studies suggest that adversarial examples tend to deviate from the underlying manifold of normal texts, whereas pre-trained masked language models can approximate the manifold of normal data. These findings inspire the exploration of masked language models for detecting textual adversarial attacks. We first introduce Masked Language Model-based Detection (MLMD), leveraging the mask and unmask operations of the masked language modeling (MLM) objective to induce the difference in m"},"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":"2504.08798","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-08T14:10:57Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"dad825f658006f64acff12360c2731c20240474c682e73b3f83e698f4a3c34f1","abstract_canon_sha256":"dc818d1a8759c54fe7efbf01fdecebb5699699fb7d5b60da08359d640d31226b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:47:59.493248Z","signature_b64":"qZqukzqd3O1CbE0dTluMW1cHJP0erhUce3xk4DiptSiumyWN9sYZ0jMp3Shlm/rNyrlp6MzUc5qNApWhE75NAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b054c28eab722a1a1991ae5e4fb973b65f2e5decd6610cbe74df8428686aa577","last_reissued_at":"2026-07-05T10:47:59.492773Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:47:59.492773Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Exploring Gradient-Guided Masked Language Model to Detect Textual Adversarial Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Leo Yu Zhang, Shengshan Hu, Shirui Pan, Xiaomei Zhang, Xufei Zheng, Yanjun Zhang, Zhaoxi Zhang","submitted_at":"2025-04-08T14:10:57Z","abstract_excerpt":"Textual adversarial examples pose serious threats to the reliability of natural language processing systems. Recent studies suggest that adversarial examples tend to deviate from the underlying manifold of normal texts, whereas pre-trained masked language models can approximate the manifold of normal data. These findings inspire the exploration of masked language models for detecting textual adversarial attacks. We first introduce Masked Language Model-based Detection (MLMD), leveraging the mask and unmask operations of the masked language modeling (MLM) objective to induce the difference in m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.08798","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/2504.08798/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":"2504.08798","created_at":"2026-07-05T10:47:59.492826+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.08798v1","created_at":"2026-07-05T10:47:59.492826+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.08798","created_at":"2026-07-05T10:47:59.492826+00:00"},{"alias_kind":"pith_short_12","alias_value":"WBKMFDVLOIVB","created_at":"2026-07-05T10:47:59.492826+00:00"},{"alias_kind":"pith_short_16","alias_value":"WBKMFDVLOIVBUGMR","created_at":"2026-07-05T10:47:59.492826+00:00"},{"alias_kind":"pith_short_8","alias_value":"WBKMFDVL","created_at":"2026-07-05T10:47:59.492826+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ","json":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ.json","graph_json":"https://pith.science/api/pith-number/WBKMFDVLOIVBUGMRVZPE7OLTWZ/graph.json","events_json":"https://pith.science/api/pith-number/WBKMFDVLOIVBUGMRVZPE7OLTWZ/events.json","paper":"https://pith.science/paper/WBKMFDVL"},"agent_actions":{"view_html":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ","download_json":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ.json","view_paper":"https://pith.science/paper/WBKMFDVL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.08798&json=true","fetch_graph":"https://pith.science/api/pith-number/WBKMFDVLOIVBUGMRVZPE7OLTWZ/graph.json","fetch_events":"https://pith.science/api/pith-number/WBKMFDVLOIVBUGMRVZPE7OLTWZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ/action/storage_attestation","attest_author":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ/action/author_attestation","sign_citation":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ/action/citation_signature","submit_replication":"https://pith.science/pith/WBKMFDVLOIVBUGMRVZPE7OLTWZ/action/replication_record"}},"created_at":"2026-07-05T10:47:59.492826+00:00","updated_at":"2026-07-05T10:47:59.492826+00:00"}