{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:UXR5NWKLRHHKQYVY3NOK6JRYIZ","short_pith_number":"pith:UXR5NWKL","canonical_record":{"source":{"id":"2408.10668","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-08-20T09:11:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"58f0202f9bfad107cd11c29abad13321e4c3e1391b25dbc61cd9820e106256e0","abstract_canon_sha256":"bd7a411758226715b839cb54aba1a76cbd435d0a49c01410b95201e072bde720"},"schema_version":"1.0"},"canonical_sha256":"a5e3d6d94b89cea862b8db5caf2638464c77372021214d20d16e37ccb7fe2fc6","source":{"kind":"arxiv","id":"2408.10668","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.10668","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"arxiv_version","alias_value":"2408.10668v3","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.10668","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"pith_short_12","alias_value":"UXR5NWKLRHHK","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"pith_short_16","alias_value":"UXR5NWKLRHHKQYVY","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"pith_short_8","alias_value":"UXR5NWKL","created_at":"2026-07-05T08:59:12Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:UXR5NWKLRHHKQYVY3NOK6JRYIZ","target":"record","payload":{"canonical_record":{"source":{"id":"2408.10668","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-08-20T09:11:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"58f0202f9bfad107cd11c29abad13321e4c3e1391b25dbc61cd9820e106256e0","abstract_canon_sha256":"bd7a411758226715b839cb54aba1a76cbd435d0a49c01410b95201e072bde720"},"schema_version":"1.0"},"canonical_sha256":"a5e3d6d94b89cea862b8db5caf2638464c77372021214d20d16e37ccb7fe2fc6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:59:12.066215Z","signature_b64":"ggLGtm2FCpntb6mwT9kRW7UyqacqnOEIrQc1u8cUvaRZgJdzfu+TxHufZJvWdeKTOwSnNmlQifX8LLNwKVdYAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a5e3d6d94b89cea862b8db5caf2638464c77372021214d20d16e37ccb7fe2fc6","last_reissued_at":"2026-07-05T08:59:12.065736Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:59:12.065736Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2408.10668","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:59:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wiFM7Qq7ca/HmIZaVO2Q3+H82i2btYCZiwz/FFinpka7Ue3XC/ueAPlheKV+QXAlsmyD3RJKcPqYHA5yWIxyAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T13:16:16.787417Z"},"content_sha256":"12a6211c1a6466ed4320e02851b86aa1c6d5daba7ab7adc0942ac4588cd97fdf","schema_version":"1.0","event_id":"sha256:12a6211c1a6466ed4320e02851b86aa1c6d5daba7ab7adc0942ac4588cd97fdf"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:UXR5NWKLRHHKQYVY3NOK6JRYIZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Probing the Safety Response Boundary of Large Language Models via Unsafe Decoding Path Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Bingzhe Wu, Haoyu Wang, Peilin Zhao, Xueqian Wang, Yatao Bian, Yongzhe Chang","submitted_at":"2024-08-20T09:11:21Z","abstract_excerpt":"Large Language Models (LLMs) are implicit troublemakers. While they provide valuable insights and assist in problem-solving, they can also potentially serve as a resource for malicious activities. Implementing safety alignment could mitigate the risk of LLMs generating harmful responses. We argue that: even when an LLM appears to successfully block harmful queries, there may still be hidden vulnerabilities that could act as ticking time bombs. To identify these underlying weaknesses, we propose to use a cost value model as both a detector and an attacker. Trained on external or self-generated "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.10668","kind":"arxiv","version":3},"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/2408.10668/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:59:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eJJgtiVn+KZ5Dss0UvsAwnooQ6t40Rz8E9Jnbd5ohdAcFgxTpGw3f5IzJSrpZM+D0UsInBXof5aFv7eyk0GjCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T13:16:16.787985Z"},"content_sha256":"d74addc123945e818402a3e91232fb9ceadfd471926b25c9803d088e10c01b2a","schema_version":"1.0","event_id":"sha256:d74addc123945e818402a3e91232fb9ceadfd471926b25c9803d088e10c01b2a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/UXR5NWKLRHHKQYVY3NOK6JRYIZ/bundle.json","state_url":"https://pith.science/pith/UXR5NWKLRHHKQYVY3NOK6JRYIZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/UXR5NWKLRHHKQYVY3NOK6JRYIZ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-08T13:16:16Z","links":{"resolver":"https://pith.science/pith/UXR5NWKLRHHKQYVY3NOK6JRYIZ","bundle":"https://pith.science/pith/UXR5NWKLRHHKQYVY3NOK6JRYIZ/bundle.json","state":"https://pith.science/pith/UXR5NWKLRHHKQYVY3NOK6JRYIZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/UXR5NWKLRHHKQYVY3NOK6JRYIZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:UXR5NWKLRHHKQYVY3NOK6JRYIZ","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":"bd7a411758226715b839cb54aba1a76cbd435d0a49c01410b95201e072bde720","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-08-20T09:11:21Z","title_canon_sha256":"58f0202f9bfad107cd11c29abad13321e4c3e1391b25dbc61cd9820e106256e0"},"schema_version":"1.0","source":{"id":"2408.10668","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.10668","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"arxiv_version","alias_value":"2408.10668v3","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.10668","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"pith_short_12","alias_value":"UXR5NWKLRHHK","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"pith_short_16","alias_value":"UXR5NWKLRHHKQYVY","created_at":"2026-07-05T08:59:12Z"},{"alias_kind":"pith_short_8","alias_value":"UXR5NWKL","created_at":"2026-07-05T08:59:12Z"}],"graph_snapshots":[{"event_id":"sha256:d74addc123945e818402a3e91232fb9ceadfd471926b25c9803d088e10c01b2a","target":"graph","created_at":"2026-07-05T08:59:12Z","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/2408.10668/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) are implicit troublemakers. While they provide valuable insights and assist in problem-solving, they can also potentially serve as a resource for malicious activities. Implementing safety alignment could mitigate the risk of LLMs generating harmful responses. We argue that: even when an LLM appears to successfully block harmful queries, there may still be hidden vulnerabilities that could act as ticking time bombs. To identify these underlying weaknesses, we propose to use a cost value model as both a detector and an attacker. Trained on external or self-generated ","authors_text":"Bingzhe Wu, Haoyu Wang, Peilin Zhao, Xueqian Wang, Yatao Bian, Yongzhe Chang","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-08-20T09:11:21Z","title":"Probing the Safety Response Boundary of Large Language Models via Unsafe Decoding Path Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.10668","kind":"arxiv","version":3},"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":"sha256:12a6211c1a6466ed4320e02851b86aa1c6d5daba7ab7adc0942ac4588cd97fdf","target":"record","created_at":"2026-07-05T08:59:12Z","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":"bd7a411758226715b839cb54aba1a76cbd435d0a49c01410b95201e072bde720","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-08-20T09:11:21Z","title_canon_sha256":"58f0202f9bfad107cd11c29abad13321e4c3e1391b25dbc61cd9820e106256e0"},"schema_version":"1.0","source":{"id":"2408.10668","kind":"arxiv","version":3}},"canonical_sha256":"a5e3d6d94b89cea862b8db5caf2638464c77372021214d20d16e37ccb7fe2fc6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a5e3d6d94b89cea862b8db5caf2638464c77372021214d20d16e37ccb7fe2fc6","first_computed_at":"2026-07-05T08:59:12.065736Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:59:12.065736Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ggLGtm2FCpntb6mwT9kRW7UyqacqnOEIrQc1u8cUvaRZgJdzfu+TxHufZJvWdeKTOwSnNmlQifX8LLNwKVdYAg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:59:12.066215Z","signed_message":"canonical_sha256_bytes"},"source_id":"2408.10668","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:12a6211c1a6466ed4320e02851b86aa1c6d5daba7ab7adc0942ac4588cd97fdf","sha256:d74addc123945e818402a3e91232fb9ceadfd471926b25c9803d088e10c01b2a"],"state_sha256":"252d77a4bae4de25020fa2f4062a98c7c0a76c693574d8fa96e94cc382accd49"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6j7LQkcH2TL2o9eYMa7WY7kETmREDPJee2r9YmdB5AUMzlUGpL3Bgtgl6RWUznJ4P6/bOLhl15c8UwnzB/u9Bg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T13:16:16.794408Z","bundle_sha256":"2e98245e14b5f8cb1e01e3c1902712f223bf6c10d5b3137d241d5885595dad23"}}