{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:B2EVLAGP7LPR2HXPY5RET6LL6U","short_pith_number":"pith:B2EVLAGP","schema_version":"1.0","canonical_sha256":"0e895580cffadf1d1eefc76249f96bf531a2437caeb0d1b52390d0ebef289365","source":{"kind":"arxiv","id":"2508.01647","version":1},"attestation_state":"computed","paper":{"title":"DUP: Detection-guided Unlearning for Backdoor Purification in Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Liangyu Chen, Man Hu, Shuai Zhao, Yahui Ding, Yanhao Jia, Yatao Yang","submitted_at":"2025-08-03T08:12:21Z","abstract_excerpt":"As backdoor attacks become more stealthy and robust, they reveal critical weaknesses in current defense strategies: detection methods often rely on coarse-grained feature statistics, and purification methods typically require full retraining or additional clean models. To address these challenges, we propose DUP (Detection-guided Unlearning for Purification), a unified framework that integrates backdoor detection with unlearning-based purification. The detector captures feature-level anomalies by jointly leveraging class-agnostic distances and inter-layer transitions. These deviations are inte"},"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":"2508.01647","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-08-03T08:12:21Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"7fc3c23b5013584cd6039b0120c5adafbb9b802ee9870bb1eacdb40a1b2de6d7","abstract_canon_sha256":"532dc7c7be907b624d47e6f8d5e06b5def6ca0e971450754fa57af358c866dd2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:47:42.878454Z","signature_b64":"SbjdmTd3C1Z4y3RA9J5S2OzhYyCi45HaxZVZVqw+9zvdKfzHEC8EYhJA+oj765blwtJbKmDSGRVUMOP5AXdqAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e895580cffadf1d1eefc76249f96bf531a2437caeb0d1b52390d0ebef289365","last_reissued_at":"2026-07-05T11:47:42.877975Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:47:42.877975Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DUP: Detection-guided Unlearning for Backdoor Purification in Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Liangyu Chen, Man Hu, Shuai Zhao, Yahui Ding, Yanhao Jia, Yatao Yang","submitted_at":"2025-08-03T08:12:21Z","abstract_excerpt":"As backdoor attacks become more stealthy and robust, they reveal critical weaknesses in current defense strategies: detection methods often rely on coarse-grained feature statistics, and purification methods typically require full retraining or additional clean models. To address these challenges, we propose DUP (Detection-guided Unlearning for Purification), a unified framework that integrates backdoor detection with unlearning-based purification. The detector captures feature-level anomalies by jointly leveraging class-agnostic distances and inter-layer transitions. These deviations are inte"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.01647","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/2508.01647/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":"2508.01647","created_at":"2026-07-05T11:47:42.878031+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.01647v1","created_at":"2026-07-05T11:47:42.878031+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.01647","created_at":"2026-07-05T11:47:42.878031+00:00"},{"alias_kind":"pith_short_12","alias_value":"B2EVLAGP7LPR","created_at":"2026-07-05T11:47:42.878031+00:00"},{"alias_kind":"pith_short_16","alias_value":"B2EVLAGP7LPR2HXP","created_at":"2026-07-05T11:47:42.878031+00:00"},{"alias_kind":"pith_short_8","alias_value":"B2EVLAGP","created_at":"2026-07-05T11:47:42.878031+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/B2EVLAGP7LPR2HXPY5RET6LL6U","json":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U.json","graph_json":"https://pith.science/api/pith-number/B2EVLAGP7LPR2HXPY5RET6LL6U/graph.json","events_json":"https://pith.science/api/pith-number/B2EVLAGP7LPR2HXPY5RET6LL6U/events.json","paper":"https://pith.science/paper/B2EVLAGP"},"agent_actions":{"view_html":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U","download_json":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U.json","view_paper":"https://pith.science/paper/B2EVLAGP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.01647&json=true","fetch_graph":"https://pith.science/api/pith-number/B2EVLAGP7LPR2HXPY5RET6LL6U/graph.json","fetch_events":"https://pith.science/api/pith-number/B2EVLAGP7LPR2HXPY5RET6LL6U/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U/action/storage_attestation","attest_author":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U/action/author_attestation","sign_citation":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U/action/citation_signature","submit_replication":"https://pith.science/pith/B2EVLAGP7LPR2HXPY5RET6LL6U/action/replication_record"}},"created_at":"2026-07-05T11:47:42.878031+00:00","updated_at":"2026-07-05T11:47:42.878031+00:00"}