{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:WL4I3RVJITAKV6CKUPVNSTKODX","short_pith_number":"pith:WL4I3RVJ","canonical_record":{"source":{"id":"2411.17693","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-26T18:58:20Z","cross_cats_sorted":[],"title_canon_sha256":"aaca6944a80103225332702b1b09a03b399f72c48712113aab8ea0d8fd5724e3","abstract_canon_sha256":"f198451c1fc1ac616b337afa4f356c879aef98a6d531974cca5bc10c5160df60"},"schema_version":"1.0"},"canonical_sha256":"b2f88dc6a944c0aaf84aa3ead94d4e1dff89c3299e8292a1248d34b9e3124afa","source":{"kind":"arxiv","id":"2411.17693","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.17693","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"arxiv_version","alias_value":"2411.17693v1","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.17693","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"pith_short_12","alias_value":"WL4I3RVJITAK","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"pith_short_16","alias_value":"WL4I3RVJITAKV6CK","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"pith_short_8","alias_value":"WL4I3RVJ","created_at":"2026-07-05T09:40:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:WL4I3RVJITAKV6CKUPVNSTKODX","target":"record","payload":{"canonical_record":{"source":{"id":"2411.17693","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-26T18:58:20Z","cross_cats_sorted":[],"title_canon_sha256":"aaca6944a80103225332702b1b09a03b399f72c48712113aab8ea0d8fd5724e3","abstract_canon_sha256":"f198451c1fc1ac616b337afa4f356c879aef98a6d531974cca5bc10c5160df60"},"schema_version":"1.0"},"canonical_sha256":"b2f88dc6a944c0aaf84aa3ead94d4e1dff89c3299e8292a1248d34b9e3124afa","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:40:50.602457Z","signature_b64":"MTnfOTjSd22LS/E/X3XUdNDIyyBpoMM1VUnkxxB4SVk+mCUdcnS/G00vCHN60T2kGPEzUnurMSxptnAhH0e/DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b2f88dc6a944c0aaf84aa3ead94d4e1dff89c3299e8292a1248d34b9e3124afa","last_reissued_at":"2026-07-05T09:40:50.601947Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:40:50.601947Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2411.17693","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":"2026-07-05T09:40:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/hswPJnK8Y7tvvO4uqA8pq11SoFu8Q0K45awY3ls0la/6dLwa1qERrntkku+uwiNUQ75t1GcJjbUPPDsGFBYDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T01:13:34.938347Z"},"content_sha256":"e385b494f9bef6c0a646b682a38cd3b2906a4cfefa292abf9abb0a367201e6b8","schema_version":"1.0","event_id":"sha256:e385b494f9bef6c0a646b682a38cd3b2906a4cfefa292abf9abb0a367201e6b8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:WL4I3RVJITAKV6CKUPVNSTKODX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Adaptive Deployment of Untrusted LLMs Reduces Distributed Threats","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Akbir Khan, Ansh Radhakrishnan, Aryan Bhatt, Buck Shlegeris, Caleb Larson, Ethan Perez, He He, Henry Sleight, Jiaxin Wen, Mrinank Sharma, Shi Feng, Vivek Hebbar","submitted_at":"2024-11-26T18:58:20Z","abstract_excerpt":"As large language models (LLMs) become increasingly capable, it is prudent to assess whether safety measures remain effective even if LLMs intentionally try to bypass them. Previous work introduced control evaluations, an adversarial framework for testing deployment strategies of untrusted models (i.e., models which might be trying to bypass safety measures). While prior work treats a single failure as unacceptable, we perform control evaluations in a \"distributed threat setting\" -- a setting where no single action is catastrophic and no single action provides overwhelming evidence of misalign"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.17693","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/2411.17693/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":"2026-07-05T09:40:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DTNQVkjTq2qiwovust1sacCRpHiO38huYxRv16+A13AQBg/zCd5EgLJiP4QPRrIrtybbQQavbV7E926ypS6bDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T01:13:34.938888Z"},"content_sha256":"a555cce307b898540cccdb89672cf1b65c9a2d86e699c54e1fe0e74f719bd161","schema_version":"1.0","event_id":"sha256:a555cce307b898540cccdb89672cf1b65c9a2d86e699c54e1fe0e74f719bd161"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WL4I3RVJITAKV6CKUPVNSTKODX/bundle.json","state_url":"https://pith.science/pith/WL4I3RVJITAKV6CKUPVNSTKODX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WL4I3RVJITAKV6CKUPVNSTKODX/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":"2026-08-11T01:13:34Z","links":{"resolver":"https://pith.science/pith/WL4I3RVJITAKV6CKUPVNSTKODX","bundle":"https://pith.science/pith/WL4I3RVJITAKV6CKUPVNSTKODX/bundle.json","state":"https://pith.science/pith/WL4I3RVJITAKV6CKUPVNSTKODX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WL4I3RVJITAKV6CKUPVNSTKODX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:WL4I3RVJITAKV6CKUPVNSTKODX","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":"f198451c1fc1ac616b337afa4f356c879aef98a6d531974cca5bc10c5160df60","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-26T18:58:20Z","title_canon_sha256":"aaca6944a80103225332702b1b09a03b399f72c48712113aab8ea0d8fd5724e3"},"schema_version":"1.0","source":{"id":"2411.17693","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.17693","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"arxiv_version","alias_value":"2411.17693v1","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.17693","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"pith_short_12","alias_value":"WL4I3RVJITAK","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"pith_short_16","alias_value":"WL4I3RVJITAKV6CK","created_at":"2026-07-05T09:40:50Z"},{"alias_kind":"pith_short_8","alias_value":"WL4I3RVJ","created_at":"2026-07-05T09:40:50Z"}],"graph_snapshots":[{"event_id":"sha256:a555cce307b898540cccdb89672cf1b65c9a2d86e699c54e1fe0e74f719bd161","target":"graph","created_at":"2026-07-05T09:40:50Z","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/2411.17693/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As large language models (LLMs) become increasingly capable, it is prudent to assess whether safety measures remain effective even if LLMs intentionally try to bypass them. Previous work introduced control evaluations, an adversarial framework for testing deployment strategies of untrusted models (i.e., models which might be trying to bypass safety measures). While prior work treats a single failure as unacceptable, we perform control evaluations in a \"distributed threat setting\" -- a setting where no single action is catastrophic and no single action provides overwhelming evidence of misalign","authors_text":"Akbir Khan, Ansh Radhakrishnan, Aryan Bhatt, Buck Shlegeris, Caleb Larson, Ethan Perez, He He, Henry Sleight, Jiaxin Wen, Mrinank Sharma, Shi Feng, Vivek Hebbar","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-26T18:58:20Z","title":"Adaptive Deployment of Untrusted LLMs Reduces Distributed Threats"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.17693","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:e385b494f9bef6c0a646b682a38cd3b2906a4cfefa292abf9abb0a367201e6b8","target":"record","created_at":"2026-07-05T09:40:50Z","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":"f198451c1fc1ac616b337afa4f356c879aef98a6d531974cca5bc10c5160df60","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-26T18:58:20Z","title_canon_sha256":"aaca6944a80103225332702b1b09a03b399f72c48712113aab8ea0d8fd5724e3"},"schema_version":"1.0","source":{"id":"2411.17693","kind":"arxiv","version":1}},"canonical_sha256":"b2f88dc6a944c0aaf84aa3ead94d4e1dff89c3299e8292a1248d34b9e3124afa","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b2f88dc6a944c0aaf84aa3ead94d4e1dff89c3299e8292a1248d34b9e3124afa","first_computed_at":"2026-07-05T09:40:50.601947Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:40:50.601947Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MTnfOTjSd22LS/E/X3XUdNDIyyBpoMM1VUnkxxB4SVk+mCUdcnS/G00vCHN60T2kGPEzUnurMSxptnAhH0e/DA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:40:50.602457Z","signed_message":"canonical_sha256_bytes"},"source_id":"2411.17693","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e385b494f9bef6c0a646b682a38cd3b2906a4cfefa292abf9abb0a367201e6b8","sha256:a555cce307b898540cccdb89672cf1b65c9a2d86e699c54e1fe0e74f719bd161"],"state_sha256":"21f0b04e87923084e473784954d4f9454261a484e6280e004bec6c6d1921abde"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"z4Cx0H6huwzND6xDgK4U1eyaWOAFQgptD9RXeQvBasvlIyoIt8IPdGv0kDC1Xf5C7zTin9H3EyP2Nr0iBTBgCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T01:13:34.942984Z","bundle_sha256":"dcbbb2442d92c97eeb1ff5b6a3b0e862304bb126eaaf1b74453d79901ef498e3"}}