{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:STG3H3RAL3CHP27SBBXMLQJUZ5","short_pith_number":"pith:STG3H3RA","schema_version":"1.0","canonical_sha256":"94cdb3ee205ec477ebf2086ec5c134cf63a4aa40ca56c45cf7ce340ec20c88b9","source":{"kind":"arxiv","id":"2608.13304","version":1},"attestation_state":"computed","paper":{"title":"Refusing Intent, Not Form: Wrapper-Based Intent-Group Supervision for LLM Safety","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Feifei Zhao, Guobin Shen, Haibo Tong, Han Shen, Ping Wu, Sicheng Shen, Yilin Zhao, Yi Zeng, Yun Luo, Yu Shi","submitted_at":"2026-08-13T14:33:15Z","abstract_excerpt":"Safety tuning can improve harmful refusal, but models may learn surface-form shortcuts: wrapped harmful prompts bypass safety, while similarly wrapped benign prompts are over-refused. We propose Wrapper-Based Intent-Form Augmentation (WIFA), an automatic intent-group augmentation method that pairs wrapped harmful examples with structurally matched wrapped benign counterexamples, requiring no external teacher or manual per-wrapper intent labels. We use WIFA as a common data layer for two complementary fine-tuning routes: WIFA-Boost, a two-stage high-safety recipe, and Anchored Group-Consistent "},"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":"2608.13304","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-08-13T14:33:15Z","cross_cats_sorted":[],"title_canon_sha256":"a3da0ae93be7b22238ae879df3adcb62dd3850118c7c6472aebf6cb324d59ecd","abstract_canon_sha256":"16075878c0a034d0d6718c2319ba1d0a4bfcd4911d4d0949c3fd08abf0925ea1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-14T01:04:10.357941Z","signature_b64":"t4h0Ha6JzUs8fNqrE+H/4JTWh39N5DM8YEZE7kRJvUKD0ghF9ZVw89tt/Ov/ZfCBE3MVm4Smd26OIltX55m0AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"94cdb3ee205ec477ebf2086ec5c134cf63a4aa40ca56c45cf7ce340ec20c88b9","last_reissued_at":"2026-08-14T01:04:10.356467Z","signature_status":"signed_v1","first_computed_at":"2026-08-14T01:04:10.356467Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Refusing Intent, Not Form: Wrapper-Based Intent-Group Supervision for LLM Safety","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Feifei Zhao, Guobin Shen, Haibo Tong, Han Shen, Ping Wu, Sicheng Shen, Yilin Zhao, Yi Zeng, Yun Luo, Yu Shi","submitted_at":"2026-08-13T14:33:15Z","abstract_excerpt":"Safety tuning can improve harmful refusal, but models may learn surface-form shortcuts: wrapped harmful prompts bypass safety, while similarly wrapped benign prompts are over-refused. We propose Wrapper-Based Intent-Form Augmentation (WIFA), an automatic intent-group augmentation method that pairs wrapped harmful examples with structurally matched wrapped benign counterexamples, requiring no external teacher or manual per-wrapper intent labels. We use WIFA as a common data layer for two complementary fine-tuning routes: WIFA-Boost, a two-stage high-safety recipe, and Anchored Group-Consistent "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.13304","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/2608.13304/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":"2608.13304","created_at":"2026-08-14T01:04:10.357253+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.13304v1","created_at":"2026-08-14T01:04:10.357253+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.13304","created_at":"2026-08-14T01:04:10.357253+00:00"},{"alias_kind":"pith_short_12","alias_value":"STG3H3RAL3CH","created_at":"2026-08-14T01:04:10.357253+00:00"},{"alias_kind":"pith_short_16","alias_value":"STG3H3RAL3CHP27S","created_at":"2026-08-14T01:04:10.357253+00:00"},{"alias_kind":"pith_short_8","alias_value":"STG3H3RA","created_at":"2026-08-14T01:04:10.357253+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/STG3H3RAL3CHP27SBBXMLQJUZ5","json":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5.json","graph_json":"https://pith.science/api/pith-number/STG3H3RAL3CHP27SBBXMLQJUZ5/graph.json","events_json":"https://pith.science/api/pith-number/STG3H3RAL3CHP27SBBXMLQJUZ5/events.json","paper":"https://pith.science/paper/STG3H3RA"},"agent_actions":{"view_html":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5","download_json":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5.json","view_paper":"https://pith.science/paper/STG3H3RA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.13304&json=true","fetch_graph":"https://pith.science/api/pith-number/STG3H3RAL3CHP27SBBXMLQJUZ5/graph.json","fetch_events":"https://pith.science/api/pith-number/STG3H3RAL3CHP27SBBXMLQJUZ5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5/action/storage_attestation","attest_author":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5/action/author_attestation","sign_citation":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5/action/citation_signature","submit_replication":"https://pith.science/pith/STG3H3RAL3CHP27SBBXMLQJUZ5/action/replication_record"}},"created_at":"2026-08-14T01:04:10.357253+00:00","updated_at":"2026-08-14T01:04:10.357253+00:00"}