{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:C7BAFTVDEVQCA3MT62FQPTP5IU","short_pith_number":"pith:C7BAFTVD","canonical_record":{"source":{"id":"2602.21198","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-24T18:55:18Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CV","cs.RO"],"title_canon_sha256":"5c27eaf0d8482cbb5b9a33d01ba2228571b72860dcd92b60b31d246ffc847fdb","abstract_canon_sha256":"25f2e722250a220f2d4d04f790431684a0c0f84d49d3875744f254df9b6129d6"},"schema_version":"1.0"},"canonical_sha256":"17c202cea32560206d93f68b07cdfd451b4587b7d8a907dad826b94c387dbf52","source":{"kind":"arxiv","id":"2602.21198","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.21198","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"arxiv_version","alias_value":"2602.21198v3","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.21198","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"pith_short_12","alias_value":"C7BAFTVDEVQC","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"pith_short_16","alias_value":"C7BAFTVDEVQCA3MT","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"pith_short_8","alias_value":"C7BAFTVD","created_at":"2026-05-26T02:04:07Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:C7BAFTVDEVQCA3MT62FQPTP5IU","target":"record","payload":{"canonical_record":{"source":{"id":"2602.21198","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-24T18:55:18Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CV","cs.RO"],"title_canon_sha256":"5c27eaf0d8482cbb5b9a33d01ba2228571b72860dcd92b60b31d246ffc847fdb","abstract_canon_sha256":"25f2e722250a220f2d4d04f790431684a0c0f84d49d3875744f254df9b6129d6"},"schema_version":"1.0"},"canonical_sha256":"17c202cea32560206d93f68b07cdfd451b4587b7d8a907dad826b94c387dbf52","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:04:07.074418Z","signature_b64":"L/oihxwPgdj5TgODlnzeZC6KhEt8XazKwhN+0dicO3LdUz9okEfIttgIWfR1+d0WgzeaMOCqaSnfat6PVguOBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"17c202cea32560206d93f68b07cdfd451b4587b7d8a907dad826b94c387dbf52","last_reissued_at":"2026-05-26T02:04:07.073593Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:04:07.073593Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2602.21198","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T02:04:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"aNBWuxiQchSl6OeEDrFp2oQyiSs4h946gt3Q8B8pEXHdE2eIM/GP+Yx4RMzeYQocHE6mQ2tsrgPW/PIFSxaWAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T00:18:51.619685Z"},"content_sha256":"c0458f000632a4477b6a95515a127b1e36581b0c4c1ee8e5ded53262448c47b6","schema_version":"1.0","event_id":"sha256:c0458f000632a4477b6a95515a127b1e36581b0c4c1ee8e5ded53262448c47b6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:C7BAFTVDEVQCA3MT62FQPTP5IU","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning from Trials and Errors: Reflective Test-Time Planning for Embodied LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Embodied LLMs improve long-horizon task performance by reflecting on failures before and after each execution at test time.","cross_cats":["cs.AI","cs.CL","cs.CV","cs.RO"],"primary_cat":"cs.LG","authors_text":"Huang Huang, Jiajun Wu, Leonidas Guibas, Li Fei-Fei, Manling Li, Yejin Choi, Yining Hong","submitted_at":"2026-02-24T18:55:18Z","abstract_excerpt":"Embodied LLMs endow robots with high-level task reasoning, but they cannot reflect on what went wrong or why, turning deployment into a sequence of independent trials where mistakes repeat rather than accumulate into experience. Drawing upon human reflective practitioners, we introduce Reflective Test-Time Planning, which integrates two modes of reflection: \\textit{reflection-in-action}, where the agent uses test-time scaling to generate and score multiple candidate actions using internal reflections before execution; and \\textit{reflection-on-action}, which uses test-time training to update b"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on our newly-designed Long-Horizon Household benchmark and MuJoCo Cupboard Fitting benchmark show significant gains over baseline models, with zero-shot generalization to photorealistic HM3D environments and real-robot experiments on a Franka Panda arm. Ablations confirm that reflection-in-action and reflection-on-action are mutually dependent, and that retrospective reflection achieves better credit assignment than step-wise external feedback at lower computational overhead.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That internal model reflections and external feedback after execution can reliably identify the causes of failures and that test-time training updates can improve the policy without instability or loss of prior capabilities.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Reflective Test-Time Planning combines pre-execution internal reflection with post-execution model updates to improve embodied LLMs on household and manipulation tasks with better long-horizon credit assignment.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Embodied LLMs improve long-horizon task performance by reflecting on failures before and after each execution at test time.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6aeef0b29f34d89eca50c8277f822abc0fecc255f08d3a7ee00447d702b814fe"},"source":{"id":"2602.21198","kind":"arxiv","version":3},"verdict":{"id":"76727724-db0d-4b3e-a149-bf9ab9226bd3","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T19:41:10.555564Z","strongest_claim":"Experiments on our newly-designed Long-Horizon Household benchmark and MuJoCo Cupboard Fitting benchmark show significant gains over baseline models, with zero-shot generalization to photorealistic HM3D environments and real-robot experiments on a Franka Panda arm. Ablations confirm that reflection-in-action and reflection-on-action are mutually dependent, and that retrospective reflection achieves better credit assignment than step-wise external feedback at lower computational overhead.","one_line_summary":"Reflective Test-Time Planning combines pre-execution internal reflection with post-execution model updates to improve embodied LLMs on household and manipulation tasks with better long-horizon credit assignment.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That internal model reflections and external feedback after execution can reliably identify the causes of failures and that test-time training updates can improve the policy without instability or loss of prior capabilities.","pith_extraction_headline":"Embodied LLMs improve long-horizon task performance by reflecting on failures before and after each execution at test time."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2602.21198/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":2,"snapshot_sha256":"122c40ea650ad326a3f0460910b38b598ec58449d188a4df0cb8e6f99d1b497a"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"76727724-db0d-4b3e-a149-bf9ab9226bd3"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T02:04:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Cnkwo3J/iWGmH5d0JO1xUrkZkqTpZ5zNb7h0eSeA4rVmJ9f7B5I/FvljwW1T6dSiZ4Mw0sYIt6SQ5KxC4nuwAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T00:18:51.620160Z"},"content_sha256":"afd3f2df66d78e5e472102b2745904b1ac1591b9e418b8a1684f45240c186856","schema_version":"1.0","event_id":"sha256:afd3f2df66d78e5e472102b2745904b1ac1591b9e418b8a1684f45240c186856"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/C7BAFTVDEVQCA3MT62FQPTP5IU/bundle.json","state_url":"https://pith.science/pith/C7BAFTVDEVQCA3MT62FQPTP5IU/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/C7BAFTVDEVQCA3MT62FQPTP5IU/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":"2026-06-02T00:18:51Z","links":{"resolver":"https://pith.science/pith/C7BAFTVDEVQCA3MT62FQPTP5IU","bundle":"https://pith.science/pith/C7BAFTVDEVQCA3MT62FQPTP5IU/bundle.json","state":"https://pith.science/pith/C7BAFTVDEVQCA3MT62FQPTP5IU/state.json","well_known_bundle":"https://pith.science/.well-known/pith/C7BAFTVDEVQCA3MT62FQPTP5IU/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:C7BAFTVDEVQCA3MT62FQPTP5IU","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":"25f2e722250a220f2d4d04f790431684a0c0f84d49d3875744f254df9b6129d6","cross_cats_sorted":["cs.AI","cs.CL","cs.CV","cs.RO"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-24T18:55:18Z","title_canon_sha256":"5c27eaf0d8482cbb5b9a33d01ba2228571b72860dcd92b60b31d246ffc847fdb"},"schema_version":"1.0","source":{"id":"2602.21198","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.21198","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"arxiv_version","alias_value":"2602.21198v3","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.21198","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"pith_short_12","alias_value":"C7BAFTVDEVQC","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"pith_short_16","alias_value":"C7BAFTVDEVQCA3MT","created_at":"2026-05-26T02:04:07Z"},{"alias_kind":"pith_short_8","alias_value":"C7BAFTVD","created_at":"2026-05-26T02:04:07Z"}],"graph_snapshots":[{"event_id":"sha256:afd3f2df66d78e5e472102b2745904b1ac1591b9e418b8a1684f45240c186856","target":"graph","created_at":"2026-05-26T02:04:07Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Experiments on our newly-designed Long-Horizon Household benchmark and MuJoCo Cupboard Fitting benchmark show significant gains over baseline models, with zero-shot generalization to photorealistic HM3D environments and real-robot experiments on a Franka Panda arm. Ablations confirm that reflection-in-action and reflection-on-action are mutually dependent, and that retrospective reflection achieves better credit assignment than step-wise external feedback at lower computational overhead."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That internal model reflections and external feedback after execution can reliably identify the causes of failures and that test-time training updates can improve the policy without instability or loss of prior capabilities."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Reflective Test-Time Planning combines pre-execution internal reflection with post-execution model updates to improve embodied LLMs on household and manipulation tasks with better long-horizon credit assignment."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Embodied LLMs improve long-horizon task performance by reflecting on failures before and after each execution at test time."}],"snapshot_sha256":"6aeef0b29f34d89eca50c8277f822abc0fecc255f08d3a7ee00447d702b814fe"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"122c40ea650ad326a3f0460910b38b598ec58449d188a4df0cb8e6f99d1b497a"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2602.21198/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Embodied LLMs endow robots with high-level task reasoning, but they cannot reflect on what went wrong or why, turning deployment into a sequence of independent trials where mistakes repeat rather than accumulate into experience. Drawing upon human reflective practitioners, we introduce Reflective Test-Time Planning, which integrates two modes of reflection: \\textit{reflection-in-action}, where the agent uses test-time scaling to generate and score multiple candidate actions using internal reflections before execution; and \\textit{reflection-on-action}, which uses test-time training to update b","authors_text":"Huang Huang, Jiajun Wu, Leonidas Guibas, Li Fei-Fei, Manling Li, Yejin Choi, Yining Hong","cross_cats":["cs.AI","cs.CL","cs.CV","cs.RO"],"headline":"Embodied LLMs improve long-horizon task performance by reflecting on failures before and after each execution at test time.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-24T18:55:18Z","title":"Learning from Trials and Errors: Reflective Test-Time Planning for Embodied LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.21198","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-15T19:41:10.555564Z","id":"76727724-db0d-4b3e-a149-bf9ab9226bd3","model_set":{"reader":"grok-4.3"},"one_line_summary":"Reflective Test-Time Planning combines pre-execution internal reflection with post-execution model updates to improve embodied LLMs on household and manipulation tasks with better long-horizon credit assignment.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Embodied LLMs improve long-horizon task performance by reflecting on failures before and after each execution at test time.","strongest_claim":"Experiments on our newly-designed Long-Horizon Household benchmark and MuJoCo Cupboard Fitting benchmark show significant gains over baseline models, with zero-shot generalization to photorealistic HM3D environments and real-robot experiments on a Franka Panda arm. Ablations confirm that reflection-in-action and reflection-on-action are mutually dependent, and that retrospective reflection achieves better credit assignment than step-wise external feedback at lower computational overhead.","weakest_assumption":"That internal model reflections and external feedback after execution can reliably identify the causes of failures and that test-time training updates can improve the policy without instability or loss of prior capabilities."}},"verdict_id":"76727724-db0d-4b3e-a149-bf9ab9226bd3"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c0458f000632a4477b6a95515a127b1e36581b0c4c1ee8e5ded53262448c47b6","target":"record","created_at":"2026-05-26T02:04:07Z","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":"25f2e722250a220f2d4d04f790431684a0c0f84d49d3875744f254df9b6129d6","cross_cats_sorted":["cs.AI","cs.CL","cs.CV","cs.RO"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-24T18:55:18Z","title_canon_sha256":"5c27eaf0d8482cbb5b9a33d01ba2228571b72860dcd92b60b31d246ffc847fdb"},"schema_version":"1.0","source":{"id":"2602.21198","kind":"arxiv","version":3}},"canonical_sha256":"17c202cea32560206d93f68b07cdfd451b4587b7d8a907dad826b94c387dbf52","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"17c202cea32560206d93f68b07cdfd451b4587b7d8a907dad826b94c387dbf52","first_computed_at":"2026-05-26T02:04:07.073593Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T02:04:07.073593Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"L/oihxwPgdj5TgODlnzeZC6KhEt8XazKwhN+0dicO3LdUz9okEfIttgIWfR1+d0WgzeaMOCqaSnfat6PVguOBg==","signature_status":"signed_v1","signed_at":"2026-05-26T02:04:07.074418Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.21198","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c0458f000632a4477b6a95515a127b1e36581b0c4c1ee8e5ded53262448c47b6","sha256:afd3f2df66d78e5e472102b2745904b1ac1591b9e418b8a1684f45240c186856"],"state_sha256":"8ca2a2d66e7da9ba16d5d52691c61dcc4d1956fd3fbea60dd621378c98f05103"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kDXr4BwXgJ7eZv6VCinyoU8lRDVfrVxIUIxSkTGOoOXYzYe5OHLB5bwf2agjS8br2PJ1WZgs81WTTgDk6iBwAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-02T00:18:51.622491Z","bundle_sha256":"9bcb92dacab6bad2799938e8748533ab31c9d13242b12e7c09be20e60b790564"}}