{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WUS5VBUK46NGOLSVOJFOTOTHHX","short_pith_number":"pith:WUS5VBUK","schema_version":"1.0","canonical_sha256":"b525da868ae79a672e55724ae9ba673de03e2741009eeeaa5e41a3127a9a73df","source":{"kind":"arxiv","id":"2607.28959","version":1},"attestation_state":"computed","paper":{"title":"Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogates","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiliang Tang, Weiyi He, Yue Xing, Yuping Lin","submitted_at":"2026-07-31T02:26:04Z","abstract_excerpt":"Adversarial training is one of the most effective defenses against adversarial attacks, yet the computational cost remains prohibitive at modern scales, especially for large language models (LLMs). While existing mitigation strategies, e.g., latent adversarial training (LAT), have been developed, they still incur a high computational cost. In this work, we comprehensively investigate computation-efficient strategies to speed up LAT from two complementary perspectives: (1) Defense-side optimization: We explore the representation fine-tuning (ReFT) within LAT, and reveal a potential issue if the"},"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":"2607.28959","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-07-31T02:26:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3ada8d34c46894b65951c8551bcb91bc7020caf6a6e38080697d603399eda8fd","abstract_canon_sha256":"527e4977fb98e693bd3da6b79843e8e6aad688c3191548d8cf153c4189e5abda"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-03T01:17:48.593377Z","signature_b64":"tZX9EBwc+76+jmDAhK5JVpxexXKiGq5yhK196ot9Rcp0bHCG1T/DEOsEV4vWKOKhbml3J6R26es4wnXbDlATCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b525da868ae79a672e55724ae9ba673de03e2741009eeeaa5e41a3127a9a73df","last_reissued_at":"2026-08-03T01:17:48.591717Z","signature_status":"signed_v1","first_computed_at":"2026-08-03T01:17:48.591717Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogates","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiliang Tang, Weiyi He, Yue Xing, Yuping Lin","submitted_at":"2026-07-31T02:26:04Z","abstract_excerpt":"Adversarial training is one of the most effective defenses against adversarial attacks, yet the computational cost remains prohibitive at modern scales, especially for large language models (LLMs). While existing mitigation strategies, e.g., latent adversarial training (LAT), have been developed, they still incur a high computational cost. In this work, we comprehensively investigate computation-efficient strategies to speed up LAT from two complementary perspectives: (1) Defense-side optimization: We explore the representation fine-tuning (ReFT) within LAT, and reveal a potential issue if the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.28959","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/2607.28959/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":"2607.28959","created_at":"2026-08-03T01:17:48.592660+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.28959v1","created_at":"2026-08-03T01:17:48.592660+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.28959","created_at":"2026-08-03T01:17:48.592660+00:00"},{"alias_kind":"pith_short_12","alias_value":"WUS5VBUK46NG","created_at":"2026-08-03T01:17:48.592660+00:00"},{"alias_kind":"pith_short_16","alias_value":"WUS5VBUK46NGOLSV","created_at":"2026-08-03T01:17:48.592660+00:00"},{"alias_kind":"pith_short_8","alias_value":"WUS5VBUK","created_at":"2026-08-03T01:17:48.592660+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/WUS5VBUK46NGOLSVOJFOTOTHHX","json":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX.json","graph_json":"https://pith.science/api/pith-number/WUS5VBUK46NGOLSVOJFOTOTHHX/graph.json","events_json":"https://pith.science/api/pith-number/WUS5VBUK46NGOLSVOJFOTOTHHX/events.json","paper":"https://pith.science/paper/WUS5VBUK"},"agent_actions":{"view_html":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX","download_json":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX.json","view_paper":"https://pith.science/paper/WUS5VBUK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.28959&json=true","fetch_graph":"https://pith.science/api/pith-number/WUS5VBUK46NGOLSVOJFOTOTHHX/graph.json","fetch_events":"https://pith.science/api/pith-number/WUS5VBUK46NGOLSVOJFOTOTHHX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX/action/storage_attestation","attest_author":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX/action/author_attestation","sign_citation":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX/action/citation_signature","submit_replication":"https://pith.science/pith/WUS5VBUK46NGOLSVOJFOTOTHHX/action/replication_record"}},"created_at":"2026-08-03T01:17:48.592660+00:00","updated_at":"2026-08-03T01:17:48.592660+00:00"}