{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:V2LGA2F52MJCXHUV4NDWQZZWZ2","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"6ae4c73a6141d82d42d01d9a1b80ca6f6a220e1d6640712307e8f46ada7391b7","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-14T07:12:44Z","title_canon_sha256":"8ae14fd25f6038631a73be57423a0dd0c0bf3076b94da84f2d97704e28ba1aee"},"schema_version":"1.0","source":{"id":"2508.10404","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.10404","created_at":"2026-07-05T11:53:55Z"},{"alias_kind":"arxiv_version","alias_value":"2508.10404v1","created_at":"2026-07-05T11:53:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.10404","created_at":"2026-07-05T11:53:55Z"},{"alias_kind":"pith_short_12","alias_value":"V2LGA2F52MJC","created_at":"2026-07-05T11:53:55Z"},{"alias_kind":"pith_short_16","alias_value":"V2LGA2F52MJCXHUV","created_at":"2026-07-05T11:53:55Z"},{"alias_kind":"pith_short_8","alias_value":"V2LGA2F5","created_at":"2026-07-05T11:53:55Z"}],"graph_snapshots":[{"event_id":"sha256:d4e07562b337864fc9a4f8feda631d6b482e364ccf27212704be8aedfedfb3b0","target":"graph","created_at":"2026-07-05T11:53:55Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2508.10404/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"With the rapid proliferation of Natural Language Processing (NLP), especially Large Language Models (LLMs), generating adversarial examples to jailbreak LLMs remains a key challenge for understanding model vulnerabilities and improving robustness. In this context, we propose a new black-box attack method that leverages the interpretability of large models. We introduce the Sparse Feature Perturbation Framework (SFPF), a novel approach for adversarial text generation that utilizes sparse autoencoders to identify and manipulate critical features in text. After using the SAE model to reconstruct ","authors_text":"Huizhen Shu, Mengqiu Tian, Qirui Wang, Xuying Li, Yuji Kosuga, Zhuo Li","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-14T07:12:44Z","title":"Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.10404","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:d9577c4163ff9b510d7af96f687c4f7fa432bb3411be910ab5eb42a91f6e3977","target":"record","created_at":"2026-07-05T11:53:55Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"6ae4c73a6141d82d42d01d9a1b80ca6f6a220e1d6640712307e8f46ada7391b7","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-14T07:12:44Z","title_canon_sha256":"8ae14fd25f6038631a73be57423a0dd0c0bf3076b94da84f2d97704e28ba1aee"},"schema_version":"1.0","source":{"id":"2508.10404","kind":"arxiv","version":1}},"canonical_sha256":"ae966068bdd3122b9e95e347686736ce807839e252400a16bcd8f3f83877d47c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ae966068bdd3122b9e95e347686736ce807839e252400a16bcd8f3f83877d47c","first_computed_at":"2026-07-05T11:53:55.615495Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:53:55.615495Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zL4dENNJq5rNV0y7kTCQ0wexFpMpMD4QtHHumbwEfLZnId7ikrh8Tg9+LlATC1CAML+L2824Tma1ekKAchdKCw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:53:55.615892Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.10404","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d9577c4163ff9b510d7af96f687c4f7fa432bb3411be910ab5eb42a91f6e3977","sha256:d4e07562b337864fc9a4f8feda631d6b482e364ccf27212704be8aedfedfb3b0"],"state_sha256":"280e6a02b4ab8ade372bb9f8938cec7f055120fbf3cb9eea412ec75ae5d0024c"}