{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:6BUC7SM75FEMG5HYDFCODJS5HM","short_pith_number":"pith:6BUC7SM7","canonical_record":{"source":{"id":"2605.29463","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T06:56:42Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a935111dab21e5dd4535f0cf36b3768391fb12e77d600c924244d561f130adf2","abstract_canon_sha256":"353d0599440322edc33b3767ffff4d734a4ab4a914446186ec0bdf59944e536f"},"schema_version":"1.0"},"canonical_sha256":"f0682fc99fe948c374f81944e1a65d3b2bfbf1b4cc28209c823bf20c53439420","source":{"kind":"arxiv","id":"2605.29463","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.29463","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"arxiv_version","alias_value":"2605.29463v1","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29463","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"pith_short_12","alias_value":"6BUC7SM75FEM","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"pith_short_16","alias_value":"6BUC7SM75FEMG5HY","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"pith_short_8","alias_value":"6BUC7SM7","created_at":"2026-05-29T01:05:40Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:6BUC7SM75FEMG5HYDFCODJS5HM","target":"record","payload":{"canonical_record":{"source":{"id":"2605.29463","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T06:56:42Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a935111dab21e5dd4535f0cf36b3768391fb12e77d600c924244d561f130adf2","abstract_canon_sha256":"353d0599440322edc33b3767ffff4d734a4ab4a914446186ec0bdf59944e536f"},"schema_version":"1.0"},"canonical_sha256":"f0682fc99fe948c374f81944e1a65d3b2bfbf1b4cc28209c823bf20c53439420","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:05:40.911371Z","signature_b64":"uxtFyJbDDerkAcaOJaVl1posXetdIt7tyDBLC8WlDV1SZ7Kds8V/rEcSVMIPaH6fHOL/ph+bn8oux1M28WFiDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f0682fc99fe948c374f81944e1a65d3b2bfbf1b4cc28209c823bf20c53439420","last_reissued_at":"2026-05-29T01:05:40.910513Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:05:40.910513Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.29463","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":"2026-05-29T01:05:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"brkyxqracuS8h6/DeO0Qb3qPH1NJT1RmMm+t+a04/zUVI7/ts4g+Qn+Wfe8q3d4wP2Ya75kjlrqH10uu6nELCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-01T00:36:42.803534Z"},"content_sha256":"779421462d94e97e4092daf1dc0651341dd6c5a661337343e1848c93a2a4e73d","schema_version":"1.0","event_id":"sha256:779421462d94e97e4092daf1dc0651341dd6c5a661337343e1848c93a2a4e73d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:6BUC7SM75FEMG5HYDFCODJS5HM","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Honest Lying: Understanding Memory Confabulation in Reflexive Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Prakhar Dixit, Sadia Kamal, Tim Oates","submitted_at":"2026-05-28T06:56:42Z","abstract_excerpt":"Reflexion-style agents rely on self-generated reflections as memory, implicitly assuming that agents can accurately diagnose their own failures.We show that this assumption can fail systematically: across ALFWorld and HumanEval, agents store confident but incorrect interpretations of the task and continue acting on them across trials,even though the environment resets to the correct task each time. We call this failure mode memory confabulation and introduce the Reflection Repetition Rate (RRR), a log-based metric that detects repeated reliance on incorrect reflective content.Using RRR, we ide"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29463","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/2605.29463/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":"2026-05-29T01:05:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+ZqpBQG/naLjmWsUEYAlkxWWGzbsIQgk433RgWZ3DxvexdCe7PNkProGr50o10jggVzGHykUp1Bktrb2niJpCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-01T00:36:42.804490Z"},"content_sha256":"371dfee7a7992b603dc437f4d84c4c96d97f9e494d5ec5113aff64cff12cc525","schema_version":"1.0","event_id":"sha256:371dfee7a7992b603dc437f4d84c4c96d97f9e494d5ec5113aff64cff12cc525"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6BUC7SM75FEMG5HYDFCODJS5HM/bundle.json","state_url":"https://pith.science/pith/6BUC7SM75FEMG5HYDFCODJS5HM/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6BUC7SM75FEMG5HYDFCODJS5HM/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":"2026-06-01T00:36:42Z","links":{"resolver":"https://pith.science/pith/6BUC7SM75FEMG5HYDFCODJS5HM","bundle":"https://pith.science/pith/6BUC7SM75FEMG5HYDFCODJS5HM/bundle.json","state":"https://pith.science/pith/6BUC7SM75FEMG5HYDFCODJS5HM/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6BUC7SM75FEMG5HYDFCODJS5HM/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:6BUC7SM75FEMG5HYDFCODJS5HM","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":"353d0599440322edc33b3767ffff4d734a4ab4a914446186ec0bdf59944e536f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T06:56:42Z","title_canon_sha256":"a935111dab21e5dd4535f0cf36b3768391fb12e77d600c924244d561f130adf2"},"schema_version":"1.0","source":{"id":"2605.29463","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.29463","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"arxiv_version","alias_value":"2605.29463v1","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29463","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"pith_short_12","alias_value":"6BUC7SM75FEM","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"pith_short_16","alias_value":"6BUC7SM75FEMG5HY","created_at":"2026-05-29T01:05:40Z"},{"alias_kind":"pith_short_8","alias_value":"6BUC7SM7","created_at":"2026-05-29T01:05:40Z"}],"graph_snapshots":[{"event_id":"sha256:371dfee7a7992b603dc437f4d84c4c96d97f9e494d5ec5113aff64cff12cc525","target":"graph","created_at":"2026-05-29T01:05:40Z","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/2605.29463/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reflexion-style agents rely on self-generated reflections as memory, implicitly assuming that agents can accurately diagnose their own failures.We show that this assumption can fail systematically: across ALFWorld and HumanEval, agents store confident but incorrect interpretations of the task and continue acting on them across trials,even though the environment resets to the correct task each time. We call this failure mode memory confabulation and introduce the Reflection Repetition Rate (RRR), a log-based metric that detects repeated reliance on incorrect reflective content.Using RRR, we ide","authors_text":"Prakhar Dixit, Sadia Kamal, Tim Oates","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T06:56:42Z","title":"Honest Lying: Understanding Memory Confabulation in Reflexive Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29463","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":"sha256:779421462d94e97e4092daf1dc0651341dd6c5a661337343e1848c93a2a4e73d","target":"record","created_at":"2026-05-29T01:05:40Z","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":"353d0599440322edc33b3767ffff4d734a4ab4a914446186ec0bdf59944e536f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T06:56:42Z","title_canon_sha256":"a935111dab21e5dd4535f0cf36b3768391fb12e77d600c924244d561f130adf2"},"schema_version":"1.0","source":{"id":"2605.29463","kind":"arxiv","version":1}},"canonical_sha256":"f0682fc99fe948c374f81944e1a65d3b2bfbf1b4cc28209c823bf20c53439420","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f0682fc99fe948c374f81944e1a65d3b2bfbf1b4cc28209c823bf20c53439420","first_computed_at":"2026-05-29T01:05:40.910513Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-29T01:05:40.910513Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"uxtFyJbDDerkAcaOJaVl1posXetdIt7tyDBLC8WlDV1SZ7Kds8V/rEcSVMIPaH6fHOL/ph+bn8oux1M28WFiDA==","signature_status":"signed_v1","signed_at":"2026-05-29T01:05:40.911371Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.29463","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:779421462d94e97e4092daf1dc0651341dd6c5a661337343e1848c93a2a4e73d","sha256:371dfee7a7992b603dc437f4d84c4c96d97f9e494d5ec5113aff64cff12cc525"],"state_sha256":"1f7c51b21f098220ba79db9ffdf74ce8ec0dc4d694f4aa256167809eb25169ce"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4UuOHnNbaP1xzbBDNbIWm4H0fpVQRnYGcNu7INRhfxLn8QIZC7T4fE/vSAKZJ19smuIX8LxjmADj/nzB7rMNDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-01T00:36:42.809303Z","bundle_sha256":"6d7160f77dd0baa780ba409e9365e14b3da1655d71e05ccd99ffc99044ee15a4"}}