{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:MZEGEATYA7AZHGPBEQDI37JEAV","short_pith_number":"pith:MZEGEATY","canonical_record":{"source":{"id":"2402.11450","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-02-18T04:16:24Z","cross_cats_sorted":[],"title_canon_sha256":"0d33fc1719599ee44b10a0f597cb4f4b4fb0dc492548dffc8427c420f7fb4aea","abstract_canon_sha256":"765acfce10ec2f3d384f06143c27638933cd97ac9476448d4e6cdfd9aa5d9665"},"schema_version":"1.0"},"canonical_sha256":"664862027807c19399e124068dfd24056cbb51020aa1e0b4d0e53fd4d15da507","source":{"kind":"arxiv","id":"2402.11450","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.11450","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"arxiv_version","alias_value":"2402.11450v2","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11450","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"pith_short_12","alias_value":"MZEGEATYA7AZ","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"pith_short_16","alias_value":"MZEGEATYA7AZHGPB","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"pith_short_8","alias_value":"MZEGEATY","created_at":"2026-07-05T08:25:32Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:MZEGEATYA7AZHGPBEQDI37JEAV","target":"record","payload":{"canonical_record":{"source":{"id":"2402.11450","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-02-18T04:16:24Z","cross_cats_sorted":[],"title_canon_sha256":"0d33fc1719599ee44b10a0f597cb4f4b4fb0dc492548dffc8427c420f7fb4aea","abstract_canon_sha256":"765acfce10ec2f3d384f06143c27638933cd97ac9476448d4e6cdfd9aa5d9665"},"schema_version":"1.0"},"canonical_sha256":"664862027807c19399e124068dfd24056cbb51020aa1e0b4d0e53fd4d15da507","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:25:32.266700Z","signature_b64":"S+P7fpQQfe+79n8OQdELqhqmlRYJ2TT4jh+ASrXGyf5mYwAIWXhhnEv5G7L/HVoyBTMOw2FVp77J7teEElXiCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"664862027807c19399e124068dfd24056cbb51020aa1e0b4d0e53fd4d15da507","last_reissued_at":"2026-07-05T08:25:32.266209Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:25:32.266209Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.11450","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:25:32Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7z37n5L0BScg2esH/lbjtuqD7spwkR5gnAVfBL5cWqwkzSJUyP35X31pP0R9DiVrfLYwFSA7O5qQ522btE2tDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T06:50:37.898618Z"},"content_sha256":"9e476710bda492f1aec0c80a45803a063f5a36083685c81da77e9128e1238850","schema_version":"1.0","event_id":"sha256:9e476710bda492f1aec0c80a45803a063f5a36083685c81da77e9128e1238850"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:MZEGEATYA7AZHGPBEQDI37JEAV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning to Learn Faster from Human Feedback with Language Model Predictive Control","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.RO","authors_text":"Adil Dostmohamed, Alex Bewley, Allen Ren, Andy Zeng, Assaf Hurwitz Michaely, Ayzaan Wahid, Baruch Tabanpour, Ben Jyenis, Brian Ichter, Carolina Parada, Chase Kew, Chuyuan Kelly Fu, Dorsa Sadigh, Dushyant Rao, Erik Frey, Fei Xia, Izhak Shafran, Jacky Liang, Jan Humplik, Jasmine Hsu, Jie Tan, Jonathan Tompson, Joss Moore, Kanishka Rao, Keerthana Gopalakrishnan, Ken Caluwaerts, Ken Oslund, Kuang-Huei Lee, Leila Takayama, Leonard Hasenclever, Maja Mataric, Maria Attarian, Maria Bauza, Marissa Giustina, Matthew Bennice, Montserrat Gonzalez Arenas, Nicolas Heess, Nikhil Joshi, Nik Stewart, Nimrod Gileadi, Peng Xu, Quan Vuong, Sean Kirmani, Ted Xiao, Tingnan Zhang, Tsang-Wei Edward Lee, Vincent Vanhoucke, Vincent Zhuang, Wenhao Yu, Ying Xu","submitted_at":"2024-02-18T04:16:24Z","abstract_excerpt":"Large language models (LLMs) have been shown to exhibit a wide range of capabilities, such as writing robot code from language commands -- enabling non-experts to direct robot behaviors, modify them based on feedback, or compose them to perform new tasks. However, these capabilities (driven by in-context learning) are limited to short-term interactions, where users' feedback remains relevant for only as long as it fits within the context size of the LLM, and can be forgotten over longer interactions. In this work, we investigate fine-tuning the robot code-writing LLMs, to remember their in-con"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11450","kind":"arxiv","version":2},"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/2402.11450/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":"2026-07-05T08:25:32Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rU6U4OKOsEvu8Sb/DFsz7xHHRl0+Sg3NwtVPJprlCaFUg6hEv9UGOz8hMWCRNturZUd4/29SW6HsJBk7+xZ9CA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T06:50:37.899287Z"},"content_sha256":"31133294ca05de0102d96681165d20060f15399481fe0b4297c3e75becfc9010","schema_version":"1.0","event_id":"sha256:31133294ca05de0102d96681165d20060f15399481fe0b4297c3e75becfc9010"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/MZEGEATYA7AZHGPBEQDI37JEAV/bundle.json","state_url":"https://pith.science/pith/MZEGEATYA7AZHGPBEQDI37JEAV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/MZEGEATYA7AZHGPBEQDI37JEAV/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":"2026-07-07T06:50:37Z","links":{"resolver":"https://pith.science/pith/MZEGEATYA7AZHGPBEQDI37JEAV","bundle":"https://pith.science/pith/MZEGEATYA7AZHGPBEQDI37JEAV/bundle.json","state":"https://pith.science/pith/MZEGEATYA7AZHGPBEQDI37JEAV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/MZEGEATYA7AZHGPBEQDI37JEAV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:MZEGEATYA7AZHGPBEQDI37JEAV","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":"765acfce10ec2f3d384f06143c27638933cd97ac9476448d4e6cdfd9aa5d9665","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-02-18T04:16:24Z","title_canon_sha256":"0d33fc1719599ee44b10a0f597cb4f4b4fb0dc492548dffc8427c420f7fb4aea"},"schema_version":"1.0","source":{"id":"2402.11450","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.11450","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"arxiv_version","alias_value":"2402.11450v2","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11450","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"pith_short_12","alias_value":"MZEGEATYA7AZ","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"pith_short_16","alias_value":"MZEGEATYA7AZHGPB","created_at":"2026-07-05T08:25:32Z"},{"alias_kind":"pith_short_8","alias_value":"MZEGEATY","created_at":"2026-07-05T08:25:32Z"}],"graph_snapshots":[{"event_id":"sha256:31133294ca05de0102d96681165d20060f15399481fe0b4297c3e75becfc9010","target":"graph","created_at":"2026-07-05T08:25:32Z","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/2402.11450/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) have been shown to exhibit a wide range of capabilities, such as writing robot code from language commands -- enabling non-experts to direct robot behaviors, modify them based on feedback, or compose them to perform new tasks. However, these capabilities (driven by in-context learning) are limited to short-term interactions, where users' feedback remains relevant for only as long as it fits within the context size of the LLM, and can be forgotten over longer interactions. In this work, we investigate fine-tuning the robot code-writing LLMs, to remember their in-con","authors_text":"Adil Dostmohamed, Alex Bewley, Allen Ren, Andy Zeng, Assaf Hurwitz Michaely, Ayzaan Wahid, Baruch Tabanpour, Ben Jyenis, Brian Ichter, Carolina Parada, Chase Kew, Chuyuan Kelly Fu, Dorsa Sadigh, Dushyant Rao, Erik Frey, Fei Xia, Izhak Shafran, Jacky Liang, Jan Humplik, Jasmine Hsu, Jie Tan, Jonathan Tompson, Joss Moore, Kanishka Rao, Keerthana Gopalakrishnan, Ken Caluwaerts, Ken Oslund, Kuang-Huei Lee, Leila Takayama, Leonard Hasenclever, Maja Mataric, Maria Attarian, Maria Bauza, Marissa Giustina, Matthew Bennice, Montserrat Gonzalez Arenas, Nicolas Heess, Nikhil Joshi, Nik Stewart, Nimrod Gileadi, Peng Xu, Quan Vuong, Sean Kirmani, Ted Xiao, Tingnan Zhang, Tsang-Wei Edward Lee, Vincent Vanhoucke, Vincent Zhuang, Wenhao Yu, Ying Xu","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-02-18T04:16:24Z","title":"Learning to Learn Faster from Human Feedback with Language Model Predictive Control"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11450","kind":"arxiv","version":2},"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":"sha256:9e476710bda492f1aec0c80a45803a063f5a36083685c81da77e9128e1238850","target":"record","created_at":"2026-07-05T08:25:32Z","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":"765acfce10ec2f3d384f06143c27638933cd97ac9476448d4e6cdfd9aa5d9665","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-02-18T04:16:24Z","title_canon_sha256":"0d33fc1719599ee44b10a0f597cb4f4b4fb0dc492548dffc8427c420f7fb4aea"},"schema_version":"1.0","source":{"id":"2402.11450","kind":"arxiv","version":2}},"canonical_sha256":"664862027807c19399e124068dfd24056cbb51020aa1e0b4d0e53fd4d15da507","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"664862027807c19399e124068dfd24056cbb51020aa1e0b4d0e53fd4d15da507","first_computed_at":"2026-07-05T08:25:32.266209Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:25:32.266209Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"S+P7fpQQfe+79n8OQdELqhqmlRYJ2TT4jh+ASrXGyf5mYwAIWXhhnEv5G7L/HVoyBTMOw2FVp77J7teEElXiCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:25:32.266700Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.11450","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9e476710bda492f1aec0c80a45803a063f5a36083685c81da77e9128e1238850","sha256:31133294ca05de0102d96681165d20060f15399481fe0b4297c3e75becfc9010"],"state_sha256":"1fc8d759b3a914f1b1042ff0e2f6e3e667b832f28f815cb9947eab124d226509"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KXOTXKUTWENpVNOADDSuAq5g0nE+VF562VjbaPfohE8yNuliG9f2sTSy8SGDF03xL+tzbo2ibT0oMPqWkukaDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T06:50:37.902598Z","bundle_sha256":"2f37e813b98f085c2f26ef149ae7a4b66abbaa1fab17e1d63e78e97191b85962"}}