{"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"}