{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:EN77RGPB7JSUOILWPYXWW3R6ZJ","short_pith_number":"pith:EN77RGPB","canonical_record":{"source":{"id":"2012.05862","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-10T18:19:48Z","cross_cats_sorted":[],"title_canon_sha256":"07b378047c3e246410c4c94cb14b2c4b3ff139c3b088cec31f826c79f8aee26c","abstract_canon_sha256":"e8a4d15e2ebad6bc2fb3aace995c1cbc9d1dafd01d1f2f3e62939fd576e87756"},"schema_version":"1.0"},"canonical_sha256":"237ff899e1fa654721767e2f6b6e3eca79c802b72ae4cfb7beac16bb7f39a0d7","source":{"kind":"arxiv","id":"2012.05862","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2012.05862","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"arxiv_version","alias_value":"2012.05862v1","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.05862","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"pith_short_12","alias_value":"EN77RGPB7JSU","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"pith_short_16","alias_value":"EN77RGPB7JSUOILW","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"pith_short_8","alias_value":"EN77RGPB","created_at":"2026-07-05T01:58:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:EN77RGPB7JSUOILWPYXWW3R6ZJ","target":"record","payload":{"canonical_record":{"source":{"id":"2012.05862","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-10T18:19:48Z","cross_cats_sorted":[],"title_canon_sha256":"07b378047c3e246410c4c94cb14b2c4b3ff139c3b088cec31f826c79f8aee26c","abstract_canon_sha256":"e8a4d15e2ebad6bc2fb3aace995c1cbc9d1dafd01d1f2f3e62939fd576e87756"},"schema_version":"1.0"},"canonical_sha256":"237ff899e1fa654721767e2f6b6e3eca79c802b72ae4cfb7beac16bb7f39a0d7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:58:37.876144Z","signature_b64":"Mnt5Yd2bKAtkvoPBJIAypvLbAssXmZnlQTWW8vvsfkvx+4s1zVacCoBQh6M5U/UdN6sXDy+bCTzMTk2pIeB7Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"237ff899e1fa654721767e2f6b6e3eca79c802b72ae4cfb7beac16bb7f39a0d7","last_reissued_at":"2026-07-05T01:58:37.875781Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:58:37.875781Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2012.05862","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":"2026-07-05T01:58:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YSeF8BMJk3hx7udqE/9o9BHuZZ5Zg2Mp1ySohTOjKvXoJ4onV685UEcMQjLXdVEF0vtRVyq7DEar4H8XYyP7Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:44:15.493395Z"},"content_sha256":"683a7eec9152ea7f99e0a6e19a93e3abb5fe676f810c5f77a6c813c0fafac33e","schema_version":"1.0","event_id":"sha256:683a7eec9152ea7f99e0a6e19a93e3abb5fe676f810c5f77a6c813c0fafac33e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:EN77RGPB7JSUOILWPYXWW3R6ZJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Understanding Learned Reward Functions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Adam Gleave, Eric J. Michaud, Stuart Russell","submitted_at":"2020-12-10T18:19:48Z","abstract_excerpt":"In many real-world tasks, it is not possible to procedurally specify an RL agent's reward function. In such cases, a reward function must instead be learned from interacting with and observing humans. However, current techniques for reward learning may fail to produce reward functions which accurately reflect user preferences. Absent significant advances in reward learning, it is thus important to be able to audit learned reward functions to verify whether they truly capture user preferences. In this paper, we investigate techniques for interpreting learned reward functions. In particular, we "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.05862","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/2012.05862/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":"2026-07-05T01:58:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HYbmmtzxTxrP8vicmn6DXauGfIxn8YasV+yNiIqYDTrUHItUtBTXP0ISDvRMXYNCjzrjMj3CFDWiE6NVRvcfCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:44:15.493912Z"},"content_sha256":"8037a7abd5924b55f086131c1ecd290de4cac230ebcd66f24b8da895c0e75f48","schema_version":"1.0","event_id":"sha256:8037a7abd5924b55f086131c1ecd290de4cac230ebcd66f24b8da895c0e75f48"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/EN77RGPB7JSUOILWPYXWW3R6ZJ/bundle.json","state_url":"https://pith.science/pith/EN77RGPB7JSUOILWPYXWW3R6ZJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/EN77RGPB7JSUOILWPYXWW3R6ZJ/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":"2026-08-09T15:44:15Z","links":{"resolver":"https://pith.science/pith/EN77RGPB7JSUOILWPYXWW3R6ZJ","bundle":"https://pith.science/pith/EN77RGPB7JSUOILWPYXWW3R6ZJ/bundle.json","state":"https://pith.science/pith/EN77RGPB7JSUOILWPYXWW3R6ZJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/EN77RGPB7JSUOILWPYXWW3R6ZJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:EN77RGPB7JSUOILWPYXWW3R6ZJ","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":"e8a4d15e2ebad6bc2fb3aace995c1cbc9d1dafd01d1f2f3e62939fd576e87756","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-10T18:19:48Z","title_canon_sha256":"07b378047c3e246410c4c94cb14b2c4b3ff139c3b088cec31f826c79f8aee26c"},"schema_version":"1.0","source":{"id":"2012.05862","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2012.05862","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"arxiv_version","alias_value":"2012.05862v1","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.05862","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"pith_short_12","alias_value":"EN77RGPB7JSU","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"pith_short_16","alias_value":"EN77RGPB7JSUOILW","created_at":"2026-07-05T01:58:37Z"},{"alias_kind":"pith_short_8","alias_value":"EN77RGPB","created_at":"2026-07-05T01:58:37Z"}],"graph_snapshots":[{"event_id":"sha256:8037a7abd5924b55f086131c1ecd290de4cac230ebcd66f24b8da895c0e75f48","target":"graph","created_at":"2026-07-05T01:58:37Z","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/2012.05862/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In many real-world tasks, it is not possible to procedurally specify an RL agent's reward function. In such cases, a reward function must instead be learned from interacting with and observing humans. However, current techniques for reward learning may fail to produce reward functions which accurately reflect user preferences. Absent significant advances in reward learning, it is thus important to be able to audit learned reward functions to verify whether they truly capture user preferences. In this paper, we investigate techniques for interpreting learned reward functions. In particular, we ","authors_text":"Adam Gleave, Eric J. Michaud, Stuart Russell","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-10T18:19:48Z","title":"Understanding Learned Reward Functions"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.05862","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":"sha256:683a7eec9152ea7f99e0a6e19a93e3abb5fe676f810c5f77a6c813c0fafac33e","target":"record","created_at":"2026-07-05T01:58:37Z","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":"e8a4d15e2ebad6bc2fb3aace995c1cbc9d1dafd01d1f2f3e62939fd576e87756","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-10T18:19:48Z","title_canon_sha256":"07b378047c3e246410c4c94cb14b2c4b3ff139c3b088cec31f826c79f8aee26c"},"schema_version":"1.0","source":{"id":"2012.05862","kind":"arxiv","version":1}},"canonical_sha256":"237ff899e1fa654721767e2f6b6e3eca79c802b72ae4cfb7beac16bb7f39a0d7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"237ff899e1fa654721767e2f6b6e3eca79c802b72ae4cfb7beac16bb7f39a0d7","first_computed_at":"2026-07-05T01:58:37.875781Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T01:58:37.875781Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Mnt5Yd2bKAtkvoPBJIAypvLbAssXmZnlQTWW8vvsfkvx+4s1zVacCoBQh6M5U/UdN6sXDy+bCTzMTk2pIeB7Cg==","signature_status":"signed_v1","signed_at":"2026-07-05T01:58:37.876144Z","signed_message":"canonical_sha256_bytes"},"source_id":"2012.05862","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:683a7eec9152ea7f99e0a6e19a93e3abb5fe676f810c5f77a6c813c0fafac33e","sha256:8037a7abd5924b55f086131c1ecd290de4cac230ebcd66f24b8da895c0e75f48"],"state_sha256":"26029c0fc8a427cef221521f6d2cd3a3950da8bb5c1b5fe0d884ee4f6e1cec4a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8aCZEgD0axMRTSBcPW7G811y6dMEtt19qvz9sUSy0d4lNCSgoQVIQm/KQlT3h+MQ6UPtWu4fu7OW85N0OQUcDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T15:44:15.499382Z","bundle_sha256":"79c522b466f5c3f093f577904be7420f744f5dc0570c699c03afdd936c365144"}}