{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:E3ASMJUMI2MW43HARTHNTORVVV","short_pith_number":"pith:E3ASMJUM","canonical_record":{"source":{"id":"2410.02173","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T03:25:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0f315bea5adadf0b40473da8cdc8dfc7e8576c6bd4ad9dd9c09b6efaaef4f8f3","abstract_canon_sha256":"647488e4f4ea339545c174a9b51ad2a94c2c97ffefecd195b801fae31ca9665d"},"schema_version":"1.0"},"canonical_sha256":"26c126268c46996e6ce08cced9ba35ad6ca9eff53573213f35da879d8cc415f2","source":{"kind":"arxiv","id":"2410.02173","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.02173","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"arxiv_version","alias_value":"2410.02173v1","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.02173","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"pith_short_12","alias_value":"E3ASMJUMI2MW","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"pith_short_16","alias_value":"E3ASMJUMI2MW43HA","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"pith_short_8","alias_value":"E3ASMJUM","created_at":"2026-07-05T09:15:17Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:E3ASMJUMI2MW43HARTHNTORVVV","target":"record","payload":{"canonical_record":{"source":{"id":"2410.02173","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T03:25:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0f315bea5adadf0b40473da8cdc8dfc7e8576c6bd4ad9dd9c09b6efaaef4f8f3","abstract_canon_sha256":"647488e4f4ea339545c174a9b51ad2a94c2c97ffefecd195b801fae31ca9665d"},"schema_version":"1.0"},"canonical_sha256":"26c126268c46996e6ce08cced9ba35ad6ca9eff53573213f35da879d8cc415f2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:15:17.246455Z","signature_b64":"+6ngG+kUUj+9SmU945I3DSLJp2TudfRbLi1qm0fU6CEPdKTqKV977GJVCZtlbEEJ5+uUfnUYJWIhiYez5YUWDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"26c126268c46996e6ce08cced9ba35ad6ca9eff53573213f35da879d8cc415f2","last_reissued_at":"2026-07-05T09:15:17.245980Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:15:17.245980Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.02173","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":"2026-07-05T09:15:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lyq/cjHq8L8eESk8+abWYVhhblsKqJxzXmIlpAh+qz+dac3NKhgbTxq5u5FUJloRIRxhmM96OiretZOuAYy3Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T05:01:31.055239Z"},"content_sha256":"3cbcff16c820005083712a517382bef8326517975637d39516f1dd3cc786f03a","schema_version":"1.0","event_id":"sha256:3cbcff16c820005083712a517382bef8326517975637d39516f1dd3cc786f03a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:E3ASMJUMI2MW43HARTHNTORVVV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Efficiently Deploying LLMs with Controlled Risk","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Matt Thomson, Michael J. Zellinger","submitted_at":"2024-10-03T03:25:56Z","abstract_excerpt":"Deploying large language models in production requires simultaneous attention to efficiency and risk control. Prior work has shown the possibility to cut costs while maintaining similar accuracy, but has neglected to focus on risk control. By contrast, here we present hierarchical chains with multi-level abstention (HCMA), which use model-intrinsic uncertainty to delegate queries along the LLM intelligence hierarchy, enabling training-free model switching based solely on black-box API calls. Our framework presents novel trade-offs between efficiency and risk. For example, deploying HCMA on MML"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.02173","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/2410.02173/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":"2026-07-05T09:15:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0ta6r23jYgUyFIUuIEioYWavk6ZTjUCKE/46Je74Vbn3nILz9WANte6bJaEi1Or1/fp7MDYH1DX1mRC7JbNMBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T05:01:31.055873Z"},"content_sha256":"7833e217eca34991b5c3e289da98a94d5fb33db271c22308f4e5c34c4f636f87","schema_version":"1.0","event_id":"sha256:7833e217eca34991b5c3e289da98a94d5fb33db271c22308f4e5c34c4f636f87"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/E3ASMJUMI2MW43HARTHNTORVVV/bundle.json","state_url":"https://pith.science/pith/E3ASMJUMI2MW43HARTHNTORVVV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/E3ASMJUMI2MW43HARTHNTORVVV/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":"2026-08-07T05:01:31Z","links":{"resolver":"https://pith.science/pith/E3ASMJUMI2MW43HARTHNTORVVV","bundle":"https://pith.science/pith/E3ASMJUMI2MW43HARTHNTORVVV/bundle.json","state":"https://pith.science/pith/E3ASMJUMI2MW43HARTHNTORVVV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/E3ASMJUMI2MW43HARTHNTORVVV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:E3ASMJUMI2MW43HARTHNTORVVV","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":"647488e4f4ea339545c174a9b51ad2a94c2c97ffefecd195b801fae31ca9665d","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T03:25:56Z","title_canon_sha256":"0f315bea5adadf0b40473da8cdc8dfc7e8576c6bd4ad9dd9c09b6efaaef4f8f3"},"schema_version":"1.0","source":{"id":"2410.02173","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.02173","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"arxiv_version","alias_value":"2410.02173v1","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.02173","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"pith_short_12","alias_value":"E3ASMJUMI2MW","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"pith_short_16","alias_value":"E3ASMJUMI2MW43HA","created_at":"2026-07-05T09:15:17Z"},{"alias_kind":"pith_short_8","alias_value":"E3ASMJUM","created_at":"2026-07-05T09:15:17Z"}],"graph_snapshots":[{"event_id":"sha256:7833e217eca34991b5c3e289da98a94d5fb33db271c22308f4e5c34c4f636f87","target":"graph","created_at":"2026-07-05T09:15:17Z","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/2410.02173/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Deploying large language models in production requires simultaneous attention to efficiency and risk control. Prior work has shown the possibility to cut costs while maintaining similar accuracy, but has neglected to focus on risk control. By contrast, here we present hierarchical chains with multi-level abstention (HCMA), which use model-intrinsic uncertainty to delegate queries along the LLM intelligence hierarchy, enabling training-free model switching based solely on black-box API calls. Our framework presents novel trade-offs between efficiency and risk. For example, deploying HCMA on MML","authors_text":"Matt Thomson, Michael J. Zellinger","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T03:25:56Z","title":"Efficiently Deploying LLMs with Controlled Risk"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.02173","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":"sha256:3cbcff16c820005083712a517382bef8326517975637d39516f1dd3cc786f03a","target":"record","created_at":"2026-07-05T09:15:17Z","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":"647488e4f4ea339545c174a9b51ad2a94c2c97ffefecd195b801fae31ca9665d","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T03:25:56Z","title_canon_sha256":"0f315bea5adadf0b40473da8cdc8dfc7e8576c6bd4ad9dd9c09b6efaaef4f8f3"},"schema_version":"1.0","source":{"id":"2410.02173","kind":"arxiv","version":1}},"canonical_sha256":"26c126268c46996e6ce08cced9ba35ad6ca9eff53573213f35da879d8cc415f2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"26c126268c46996e6ce08cced9ba35ad6ca9eff53573213f35da879d8cc415f2","first_computed_at":"2026-07-05T09:15:17.245980Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:15:17.245980Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"+6ngG+kUUj+9SmU945I3DSLJp2TudfRbLi1qm0fU6CEPdKTqKV977GJVCZtlbEEJ5+uUfnUYJWIhiYez5YUWDg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:15:17.246455Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.02173","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3cbcff16c820005083712a517382bef8326517975637d39516f1dd3cc786f03a","sha256:7833e217eca34991b5c3e289da98a94d5fb33db271c22308f4e5c34c4f636f87"],"state_sha256":"504c487e66aec9516c0995d3bf399eac1b3a6540d04193399ad5e01971ba7ca2"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XtcpRy+AkhIX11k65edOK20pXSYL8D6aeTjHv/oUMTP+tWPRqdukhbW1s+9kmdozkTqJataYZTHe0lTCpr/0Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T05:01:31.060790Z","bundle_sha256":"884d10dcb1cdd9202a38df44ff91728a959d47a069f5e2a723ce052c96a20b67"}}