{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:4W7RDV4A6N4N3V4J3L4EE5PEEC","short_pith_number":"pith:4W7RDV4A","canonical_record":{"source":{"id":"2401.17633","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-31T07:26:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a722110d8ca511144eab65d67708b3cbe0bb6112950e65b71abec1c9338e029a","abstract_canon_sha256":"00f72dd64e8d3bdc4c5c5bf87cd051cdf4722df60a3e9ad3f277c83048eb4b6f"},"schema_version":"1.0"},"canonical_sha256":"e5bf11d780f378ddd789daf84275e420930b8682b4e1c93855bdb7ee888ca4cc","source":{"kind":"arxiv","id":"2401.17633","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.17633","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"arxiv_version","alias_value":"2401.17633v1","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.17633","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"pith_short_12","alias_value":"4W7RDV4A6N4N","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"pith_short_16","alias_value":"4W7RDV4A6N4N3V4J","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"pith_short_8","alias_value":"4W7RDV4A","created_at":"2026-07-05T07:39:39Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:4W7RDV4A6N4N3V4J3L4EE5PEEC","target":"record","payload":{"canonical_record":{"source":{"id":"2401.17633","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-31T07:26:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a722110d8ca511144eab65d67708b3cbe0bb6112950e65b71abec1c9338e029a","abstract_canon_sha256":"00f72dd64e8d3bdc4c5c5bf87cd051cdf4722df60a3e9ad3f277c83048eb4b6f"},"schema_version":"1.0"},"canonical_sha256":"e5bf11d780f378ddd789daf84275e420930b8682b4e1c93855bdb7ee888ca4cc","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:39:39.729091Z","signature_b64":"xmIR3apjqEsDX6XlEx/ilEP+XoFSWcZdm8u8nhdgBecBkqvIVw+uVguFKSfCDiDqu1DZ25SIRbC1Wrwy8Fb+Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e5bf11d780f378ddd789daf84275e420930b8682b4e1c93855bdb7ee888ca4cc","last_reissued_at":"2026-07-05T07:39:39.728607Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:39:39.728607Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2401.17633","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:39:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZUZx/dhpjWwAiUNbUXY2anhwACCmP1ROYmo9oE6HElzDVLJFNgpiAkNcm5ntWEHKpS+ZkpWURR+ZPv/lnap6CA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T19:40:34.468620Z"},"content_sha256":"8090b57a0b942d353e1c4ecd4dd2123bdf30a589ec63f82d6cde3ed81fcd4560","schema_version":"1.0","event_id":"sha256:8090b57a0b942d353e1c4ecd4dd2123bdf30a589ec63f82d6cde3ed81fcd4560"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:4W7RDV4A6N4N3V4J3L4EE5PEEC","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Navigating the OverKill in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chenyu Shi, Dahua Lin, Qiming Ge, Qi Zhang, Songyang Gao, Tao Gui, Xianjun Yang, Xiao Wang, Xuanjing Huang, Xun Zhao","submitted_at":"2024-01-31T07:26:47Z","abstract_excerpt":"Large language models are meticulously aligned to be both helpful and harmless. However, recent research points to a potential overkill which means models may refuse to answer benign queries. In this paper, we investigate the factors for overkill by exploring how models handle and determine the safety of queries. Our findings reveal the presence of shortcuts within models, leading to an over-attention of harmful words like 'kill' and prompts emphasizing safety will exacerbate overkill. Based on these insights, we introduce Self-Contrastive Decoding (Self-CD), a training-free and model-agnostic"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.17633","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/2401.17633/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":"2026-07-05T07:39:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Omo7AyVbLe9eZ8FVrPIowX0zOBtrmXOfzChLgGaIciwVD13A5hNwZRhV20ycOsuyjKxO+IL6sY2EoNzaO8ZuBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T19:40:34.469132Z"},"content_sha256":"8e87cb2e47e152090349d6503fa155b3b8eb741a771830ef4f0471c3eacfa7ad","schema_version":"1.0","event_id":"sha256:8e87cb2e47e152090349d6503fa155b3b8eb741a771830ef4f0471c3eacfa7ad"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4W7RDV4A6N4N3V4J3L4EE5PEEC/bundle.json","state_url":"https://pith.science/pith/4W7RDV4A6N4N3V4J3L4EE5PEEC/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4W7RDV4A6N4N3V4J3L4EE5PEEC/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":"2026-08-10T19:40:34Z","links":{"resolver":"https://pith.science/pith/4W7RDV4A6N4N3V4J3L4EE5PEEC","bundle":"https://pith.science/pith/4W7RDV4A6N4N3V4J3L4EE5PEEC/bundle.json","state":"https://pith.science/pith/4W7RDV4A6N4N3V4J3L4EE5PEEC/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4W7RDV4A6N4N3V4J3L4EE5PEEC/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:4W7RDV4A6N4N3V4J3L4EE5PEEC","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":"00f72dd64e8d3bdc4c5c5bf87cd051cdf4722df60a3e9ad3f277c83048eb4b6f","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-31T07:26:47Z","title_canon_sha256":"a722110d8ca511144eab65d67708b3cbe0bb6112950e65b71abec1c9338e029a"},"schema_version":"1.0","source":{"id":"2401.17633","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.17633","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"arxiv_version","alias_value":"2401.17633v1","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.17633","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"pith_short_12","alias_value":"4W7RDV4A6N4N","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"pith_short_16","alias_value":"4W7RDV4A6N4N3V4J","created_at":"2026-07-05T07:39:39Z"},{"alias_kind":"pith_short_8","alias_value":"4W7RDV4A","created_at":"2026-07-05T07:39:39Z"}],"graph_snapshots":[{"event_id":"sha256:8e87cb2e47e152090349d6503fa155b3b8eb741a771830ef4f0471c3eacfa7ad","target":"graph","created_at":"2026-07-05T07:39:39Z","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/2401.17633/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models are meticulously aligned to be both helpful and harmless. However, recent research points to a potential overkill which means models may refuse to answer benign queries. In this paper, we investigate the factors for overkill by exploring how models handle and determine the safety of queries. Our findings reveal the presence of shortcuts within models, leading to an over-attention of harmful words like 'kill' and prompts emphasizing safety will exacerbate overkill. Based on these insights, we introduce Self-Contrastive Decoding (Self-CD), a training-free and model-agnostic","authors_text":"Chenyu Shi, Dahua Lin, Qiming Ge, Qi Zhang, Songyang Gao, Tao Gui, Xianjun Yang, Xiao Wang, Xuanjing Huang, Xun Zhao","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-31T07:26:47Z","title":"Navigating the OverKill in Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.17633","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":"sha256:8090b57a0b942d353e1c4ecd4dd2123bdf30a589ec63f82d6cde3ed81fcd4560","target":"record","created_at":"2026-07-05T07:39:39Z","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":"00f72dd64e8d3bdc4c5c5bf87cd051cdf4722df60a3e9ad3f277c83048eb4b6f","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-31T07:26:47Z","title_canon_sha256":"a722110d8ca511144eab65d67708b3cbe0bb6112950e65b71abec1c9338e029a"},"schema_version":"1.0","source":{"id":"2401.17633","kind":"arxiv","version":1}},"canonical_sha256":"e5bf11d780f378ddd789daf84275e420930b8682b4e1c93855bdb7ee888ca4cc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e5bf11d780f378ddd789daf84275e420930b8682b4e1c93855bdb7ee888ca4cc","first_computed_at":"2026-07-05T07:39:39.728607Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:39:39.728607Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"xmIR3apjqEsDX6XlEx/ilEP+XoFSWcZdm8u8nhdgBecBkqvIVw+uVguFKSfCDiDqu1DZ25SIRbC1Wrwy8Fb+Aw==","signature_status":"signed_v1","signed_at":"2026-07-05T07:39:39.729091Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.17633","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8090b57a0b942d353e1c4ecd4dd2123bdf30a589ec63f82d6cde3ed81fcd4560","sha256:8e87cb2e47e152090349d6503fa155b3b8eb741a771830ef4f0471c3eacfa7ad"],"state_sha256":"05c0193759af361eaeeb3db3c66fc6c172868c642b737e741966d8ec99bb10c0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2DlX1GHhWC7R3Fp7mbZlge33iuIVfV1+sN7cDu6HIrDfqh22CcCvsMHtA6TINRPpazgdwcUNrEZFx7hBMIWqCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T19:40:34.473954Z","bundle_sha256":"ccf4c6d8f8bef7c320f3501d80a405df79457b62c97dde39022907c5b2927233"}}