{"as_of":"2026-08-08T20:20:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:0b444b939e839dd18a2d187c75ff3671adbbac62839b3bf49dd45225ec8b4838","coverage":[{"denominator":26,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":26,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-03T12:23:58.895544Z","state":"measured"},{"denominator":26,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":26,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-08T06:32:00.761636+00:00","state":"measured"},{"denominator":0,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"cited_works","source_observed_at":null,"state":"measured"}],"external_citation_measurements":[],"inbound":[],"links":{"evidence":"/evidence","html":"/paper/2607.29172/citation-record","integrity":"/paper/2607.29172/integrity","json":"/paper/2607.29172/citation-record.json","paper":"/paper/2607.29172"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:56.547649Z","title":"Open X-embodiment: Robotic learning datasets and RT-X models","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:56.547649Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:6a69fa9d11b15ff129971d3b22aff0411ab3ac90bc2b0ac8df5480d27b352ce8","observation_id":"773c576b-a3fc-4505-90fc-e2cb5ffce8a4","resolution":{"observed_at":"2026-08-03T12:23:56.547649Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2503.20020","last_updated":"2025-03-25T19:02:56Z","snapshot_observed_at":"2026-08-02T07:15:58.798604Z","submitted_at":"2025-03-25T19:02:56Z","title":"Gemini Robotics: Bringing AI into the Physical World","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2503.20020","snapshot_observed_at":"2026-08-03T12:23:56.625409Z","title":"Gemini robotics: Bringing ai into the physical world.arXiv preprint arXiv:2503.20020, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:56.625409Z"},"links":{"cited_paper":"/paper/2503.20020","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:4ef92af81795100564ad244ddfea4adcda77ec2f0f9ffe7d47aea934de12781f","observation_id":"65c3a603-7d4a-4231-a03c-5aef55e1c933","resolution":{"observed_at":"2026-08-03T12:23:56.625409Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2503.14734","last_updated":"2025-03-27T02:52:43Z","snapshot_observed_at":"2026-08-02T04:15:31.100670Z","submitted_at":"2025-03-18T21:06:21Z","title":"GR00T N1: An Open Foundation Model for Generalist Humanoid Robots","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2503.14734","snapshot_observed_at":"2026-08-03T12:23:56.741556Z","title":"Gr00t n1: An open foundation model for generalist humanoid robots","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:56.741556Z"},"links":{"cited_paper":"/paper/2503.14734","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:6bbf3c7ff718e3323b4db309e802d9f9f75eee8ef95aaa1594969355170b687f","observation_id":"fd8e3135-67ea-4350-8dd4-fb86c15f9c39","resolution":{"observed_at":"2026-08-03T12:23:56.741556Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:56.862344Z","title":"arXiv preprint arXiv:2603.12263, 2026","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:56.862344Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:7d06c40389135394e7344389e8bb2224eb8839b5787f76161cc914508e038c17","observation_id":"a0df2964-dcc5-497d-8bc2-6cbf1bc255e1","resolution":{"observed_at":"2026-08-03T12:23:56.862344Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2511.14759","last_updated":"2025-11-19T04:34:49Z","snapshot_observed_at":"2026-07-06T22:36:13.287872Z","submitted_at":"2025-11-18T18:58:55Z","title":"$\\pi^{*}_{0.6}$: a VLA That Learns From Experience","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2511.14759","snapshot_observed_at":"2026-08-03T12:23:56.987236Z","title":"arXiv preprint arXiv:2511.14759, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:56.987236Z"},"links":{"cited_paper":"/paper/2511.14759","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:0c9e7617c4998c979b5dfd1070a7dfc05f8c62892a35699d4ae68bd5cf743a9c","observation_id":"8d359516-cd98-48e2-b5bb-c420f2e0f631","resolution":{"observed_at":"2026-08-03T12:23:56.987236Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.067804Z","title":"Helix: A vision-language-action model for generalist humanoid control, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.067804Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:265b7a818d271d17a4246fae63da885091cb8876e7e4bc614b7a6b90954660d5","observation_id":"e49b73b6-e935-4667-a598-c69e96947543","resolution":{"observed_at":"2026-08-03T12:23:57.067804Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.175879Z","title":"Fine-tuning – openai api documentation, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.175879Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:7a01bbba67d185a312bf9b4630cce7eb3982ca025e94405e97ef9435fabfedb8","observation_id":"fb3613e4-794a-428f-99e7-3f2830dfbcdb","resolution":{"observed_at":"2026-08-03T12:23:57.175879Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.256490Z","title":"About supervised fine-tuning for Gemini models – Vertex AI documentation, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.256490Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:d5eb67c3102eb04f11179a026a22dde5d56d2f5b13ff32e03fbc88bf04899359","observation_id":"40035f6d-c222-4aa2-bb6f-ded0e98479d0","resolution":{"observed_at":"2026-08-03T12:23:57.256490Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.362980Z","title":"Position: Good embodied reward models need bad behavior data","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.362980Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:05fc8a28247b03b86019af127edab6d6616cf913362014e8e2371ae08617960c","observation_id":"d04864b2-c139-4d51-bb3f-08d7fce4ed61","resolution":{"observed_at":"2026-08-03T12:23:57.362980Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.435430Z","title":"Gemini robotics on-device, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.435430Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:6f78d9896d6fa75c3148df2955c54bc9cbe50b91596c06a5cdb4d2638f431b0b","observation_id":"d6475949-fbd1-4b3c-bcf8-27107961ea74","resolution":{"observed_at":"2026-08-03T12:23:57.435430Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.553563Z","title":"The physical intelligence layer, February 2026","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.553563Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:fe9546283e98b06de2524564ecaa165827b085bd28a09cdfc4199e80985c2429","observation_id":"655bf451-c6b0-4f17-be35-a5f6b9bfd29f","resolution":{"observed_at":"2026-08-03T12:23:57.553563Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2510.08807","last_updated":"2026-07-04T03:31:36Z","snapshot_observed_at":"2026-08-04T10:44:29.379805Z","submitted_at":"2025-10-09T20:43:27Z","title":"Humanoid Everyday: A Comprehensive Robotic Dataset for Open-World Humanoid Manipulation","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2510.08807","snapshot_observed_at":"2026-08-03T12:23:57.674445Z","title":"Humanoid everyday: A comprehensive robotic dataset for open-world humanoid manipulation.arXiv preprint arXiv:2510.08807, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.674445Z"},"links":{"cited_paper":"/paper/2510.08807","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:f71848b266d16b0897bf0a96b414904221345a2fdf5cb326c8ef579aad66458f","observation_id":"4619e291-3a73-49f5-a0a4-5e03fea7c37a","resolution":{"observed_at":"2026-08-03T12:23:57.674445Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2604.02523","last_updated":"2026-04-02T21:23:08Z","snapshot_observed_at":"2026-07-06T22:51:53.458444Z","submitted_at":"2026-04-02T21:23:08Z","title":"Tune to Learn: How Controller Gains Shape Robot Policy Learning","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2604.02523","snapshot_observed_at":"2026-08-03T12:23:57.763265Z","title":"Tune to learn: How controller gains shape robot policy learning.arXiv preprint arXiv:2604.02523, 2026","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.763265Z"},"links":{"cited_paper":"/paper/2604.02523","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:ab72eb723cab4d66f2d619ea7ee988888542de810ec24f9a2582d1e414a50f94","observation_id":"88672ccc-dfd7-45d3-9c08-0e56ac783455","resolution":{"observed_at":"2026-08-03T12:23:57.763265Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.842549Z","title":"Self-improving vision-language-action models with data generation via residual RL.arXiv preprint arXiv:2511.00091, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.842549Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:5e5003e44334355f3ff258972c2a1f27244ecd9d4d8548b646fdb81105418796","observation_id":"cc675e91-f442-4839-87f9-3ea537ec12b5","resolution":{"observed_at":"2026-08-03T12:23:57.842549Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:57.894935Z","title":"Robo-Dopamine: General process reward modeling for high-precision robotic manipula- tion.arXiv preprint arXiv:2512.23703, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.894935Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:66823442e4f998a48d1750228455dfb04466d47947297c7028952d7453568705","observation_id":"f89294bc-e3f1-4137-926f-83e2f44ee253","resolution":{"observed_at":"2026-08-03T12:23:57.894935Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2506.15799","last_updated":"2025-06-25T19:09:52Z","snapshot_observed_at":"2026-08-08T08:13:32.220639Z","submitted_at":"2025-06-18T18:35:57Z","title":"Steering Your Diffusion Policy with Latent Space Reinforcement Learning","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2506.15799","snapshot_observed_at":"2026-08-03T12:23:57.989564Z","title":"Steering your diffusion policy with latent space reinforcement learning.arXiv preprint arXiv:2506.15799, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:57.989564Z"},"links":{"cited_paper":"/paper/2506.15799","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:8852303176378c82457bdd96729ff2f4a68740c338bcd7b87ba0e7bf4dcba4b7","observation_id":"35b3d4f0-b520-4f03-bace-524f0788b17d","resolution":{"observed_at":"2026-08-03T12:23:57.989564Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:58.111620Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.111620Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:4607d6c140d5d10d41af3c9641a0488e2627fb839b43e3a1cd4267b80658b1c4","observation_id":"25f77e05-9e7c-4b13-9e83-03ea8de0acbf","resolution":{"observed_at":"2026-08-03T12:23:58.111620Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1707.06347","last_updated":"2017-08-28T09:20:06Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2017-07-20T02:32:33Z","title":"Proximal Policy Optimization Algorithms","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1707.06347","snapshot_observed_at":"2026-08-03T12:23:58.169775Z","title":"Proximal policy opti- mization algorithms.arXiv preprint arXiv:1707.06347, 2017","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.169775Z"},"links":{"cited_paper":"/paper/1707.06347","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:29342b76dd6b1848d915282e957171f35c039f18b9424e97249d09526058ba85","observation_id":"bcf09b61-852a-4e03-86ed-34c9c80716ea","resolution":{"observed_at":"2026-08-03T12:23:58.169775Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1910.00177","last_updated":"2019-10-07T20:23:21Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2019-10-01T02:23:38Z","title":"Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learning","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1910.00177","snapshot_observed_at":"2026-08-03T12:23:58.216903Z","title":"Advantage-weighted regression: Simple and scalable off-policy reinforcement learning.arXiv preprint arXiv:1910.00177, 2019","venue":null,"work_id":null,"year":1910},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.216903Z"},"links":{"cited_paper":"/paper/1910.00177","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:efc547e0380b8da9caa9b5e84acb0f4c9a5b2466247bc020a397504597308d25","observation_id":"2899bc85-48f0-4de4-9012-2e71820cbbce","resolution":{"observed_at":"2026-08-03T12:23:58.216903Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:58.246760Z","title":"Vision language models are in-context value learners","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.246760Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:a14c383103c42bbc71358c2cebf1cf4fb824c689eb900b9eb156dca28a0eced7","observation_id":"517e405e-1069-4a8c-9c9b-2e89e46e8806","resolution":{"observed_at":"2026-08-03T12:23:58.246760Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2412.04835","last_updated":"2024-12-06T08:04:02Z","snapshot_observed_at":"2026-08-04T19:14:18.217627Z","submitted_at":"2024-12-06T08:04:02Z","title":"Maximizing Alignment with Minimal Feedback: Efficiently Learning Rewards for Visuomotor Robot Policy Alignment","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2412.04835","snapshot_observed_at":"2026-08-03T12:23:58.275386Z","title":"Maximizing alignment with minimal feedback: Efficiently learning rewards for visuomotor robot policy alignment","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.275386Z"},"links":{"cited_paper":"/paper/2412.04835","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:756810b7dcc052355f1354b830cc7fb9c1f93efbf0cc1f52a81ab141565e27bc","observation_id":"96684f18-60ad-4f5b-9462-9273155a96e9","resolution":{"observed_at":"2026-08-03T12:23:58.275386Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:58.383990Z","title":"XR-Teleoperate: An open-source teleoperation framework and data collection toolkit for embodied intelligence","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.383990Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:9a468995d478155a534b3b5d32579f0bd61b1a1aec18264e114d00880c4171a5","observation_id":"f54fdfef-98ea-4187-a4a2-0671e29d2b07","resolution":{"observed_at":"2026-08-03T12:23:58.383990Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2407.07726","last_updated":"2024-10-10T17:28:23Z","snapshot_observed_at":"2026-08-08T07:16:45.596308Z","submitted_at":"2024-07-10T14:57:46Z","title":"PaliGemma: A versatile 3B VLM for transfer","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2407.07726","snapshot_observed_at":"2026-08-03T12:23:58.523242Z","title":"PaliGemma: A versatile 3B VLM for transfer.arXiv preprint arXiv:2407.07726, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.523242Z"},"links":{"cited_paper":"/paper/2407.07726","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:f68de4b6c950879980906027e6e9ce106aca4515efe70734dd0e7b0d56aa2263","observation_id":"9b5776b2-ae95-4a8f-a482-c18863766235","resolution":{"observed_at":"2026-08-03T12:23:58.523242Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T12:23:58.685514Z","title":"FiLM:Visualreasoning with a general conditioning layer","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.685514Z"},"links":{"citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:7f1956fce42f335ba97645f1fd6e829bf25163e9682dbcb99496e2c5303b6d81","observation_id":"307e5376-72af-41fe-8c4e-3394655284ed","resolution":{"observed_at":"2026-08-03T12:23:58.685514Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2508.10104","last_updated":"2025-08-13T18:00:55Z","snapshot_observed_at":"2026-07-06T22:12:35.584339Z","submitted_at":"2025-08-13T18:00:55Z","title":"DINOv3","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2508.10104","snapshot_observed_at":"2026-08-03T12:23:58.809848Z","title":"Dinov3.arXivpreprintarXiv:2508.10104, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.809848Z"},"links":{"cited_paper":"/paper/2508.10104","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:de929eba91f87f293c48cd8e863e4124e8f6104ee1b0d482fc6c0b08a5f3091e","observation_id":"b583184f-f1eb-48d6-8f31-f0a55a93fabd","resolution":{"observed_at":"2026-08-03T12:23:58.809848Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2505.23458","last_updated":"2025-05-29T14:06:50Z","snapshot_observed_at":"2026-08-08T10:39:46.769874Z","submitted_at":"2025-05-29T14:06:50Z","title":"Diffusion Guidance Is a Controllable Policy Improvement Operator","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2505.23458","snapshot_observed_at":"2026-08-03T12:23:58.895544Z","title":"frame_index","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning","version":1},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-03T12:23:58.895544Z"},"links":{"cited_paper":"/paper/2505.23458","citing_paper":"/paper/2607.29172"},"observation_digest":"sha256:b5551eecede6fc6df9a27e94567f05f9b437ed40666ed2b82ff82842ae9be34f","observation_id":"9f3b3865-8852-4d00-93fa-510cccdb0c58","resolution":{"observed_at":"2026-08-03T12:23:58.895544Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2607.29172","last_updated":"2026-07-31T08:50:25Z","latest_version":1,"primary_category":"cs.RO","snapshot_observed_at":"2026-08-08T10:40:09.041344Z","submitted_at":"2026-07-31T08:50:25Z","title":"CLIFT: Turning Gemini Robotics On-Device into Humanoid Specialists via Non-Invasive Closed-Loop Iterative Fine-Tuning"},"reference_resolution":{"displayed":26,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":26,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":26},"refusal":"A citation records a reference. It does not transfer a finding from one paper to another.","schema":"pith.paper-citation-record.v1","standing_sources":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"thesis":"As of 8 August 2026, this Paper Citation Record lists 26 of 26 outbound references and 0 inbound Pith citation observations for arXiv:2607.29172."}