{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:5MQZFVEFUTKRRCVJIB2PNQ56KB","short_pith_number":"pith:5MQZFVEF","schema_version":"1.0","canonical_sha256":"eb2192d485a4d5188aa94074f6c3be5045135668917b21236a0ecca40d270964","source":{"kind":"arxiv","id":"2303.07280","version":1},"attestation_state":"computed","paper":{"title":"Vision-Language Models as Success Detectors","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Akhil Raju, Felix Hill, Jessica Landon, Ksenia Konyushkova, Misha Denil, Nando de Freitas, Serkan Cabi, Yuqing Du","submitted_at":"2023-03-13T16:54:11Z","abstract_excerpt":"Detecting successful behaviour is crucial for training intelligent agents. As such, generalisable reward models are a prerequisite for agents that can learn to generalise their behaviour. In this work we focus on developing robust success detectors that leverage large, pretrained vision-language models (Flamingo, Alayrac et al. (2022)) and human reward annotations. Concretely, we treat success detection as a visual question answering (VQA) problem, denoted SuccessVQA. We study success detection across three vastly different domains: (i) interactive language-conditioned agents in a simulated ho"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2303.07280","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-03-13T16:54:11Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"34ba897f533d6700fa067dd0e61e8ec81e16ec1bc4a8a02f7bdbd789bba80964","abstract_canon_sha256":"68e99364031898fdc622710efaadcddd465d640a7f0bbf48e0d5a4b3560077da"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:50:34.502410Z","signature_b64":"QQJxfPNcWoBA4Aq1hktCgfUSvcaA1WVhUe5u8OhSHhuMFyZVwPcICP9jcfa8gnHf62mBdHwiITh0LdTSeE3hAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eb2192d485a4d5188aa94074f6c3be5045135668917b21236a0ecca40d270964","last_reissued_at":"2026-07-05T05:50:34.501964Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:50:34.501964Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Vision-Language Models as Success Detectors","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Akhil Raju, Felix Hill, Jessica Landon, Ksenia Konyushkova, Misha Denil, Nando de Freitas, Serkan Cabi, Yuqing Du","submitted_at":"2023-03-13T16:54:11Z","abstract_excerpt":"Detecting successful behaviour is crucial for training intelligent agents. As such, generalisable reward models are a prerequisite for agents that can learn to generalise their behaviour. In this work we focus on developing robust success detectors that leverage large, pretrained vision-language models (Flamingo, Alayrac et al. (2022)) and human reward annotations. Concretely, we treat success detection as a visual question answering (VQA) problem, denoted SuccessVQA. We study success detection across three vastly different domains: (i) interactive language-conditioned agents in a simulated ho"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.07280","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/2303.07280/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2303.07280","created_at":"2026-07-05T05:50:34.502021+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.07280v1","created_at":"2026-07-05T05:50:34.502021+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.07280","created_at":"2026-07-05T05:50:34.502021+00:00"},{"alias_kind":"pith_short_12","alias_value":"5MQZFVEFUTKR","created_at":"2026-07-05T05:50:34.502021+00:00"},{"alias_kind":"pith_short_16","alias_value":"5MQZFVEFUTKRRCVJ","created_at":"2026-07-05T05:50:34.502021+00:00"},{"alias_kind":"pith_short_8","alias_value":"5MQZFVEF","created_at":"2026-07-05T05:50:34.502021+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23640","citing_title":"Learning Process Rewards via Success Visitation Matching for Efficient RL","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21572","citing_title":"Robot Critics that Sweat the Small Stuff","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10267","citing_title":"What Matters in Orchestrating Robot Policies: A Systematic Study of Hierarchical VLA Agents","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05660","citing_title":"Safe Embodied AI for Long-horizon Tasks: A Cross-layer Analysis of Robotic Manipulation","ref_index":296,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03134","citing_title":"How Visible Are Silent Manipulation Failures? An Observability Study of False-Success Detection in Simulated Robot Episodes","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2510.19268","citing_title":"Hierarchical DLO Routing with Reinforcement Learning and In-Context Vision-language Models","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2309.16797","citing_title":"Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution","ref_index":98,"is_internal_anchor":false},{"citing_arxiv_id":"2503.22020","citing_title":"CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2510.10125","citing_title":"Ctrl-World: A Controllable Generative World Model for Robot Manipulation","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2412.10345","citing_title":"TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11951","citing_title":"From Reaction to Anticipation: Proactive Failure Recovery through Agentic Task Graph for Robotic Manipulation","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07034","citing_title":"KITE: Keyframe-Indexed Tokenized Evidence for VLM-Based Robot Failure Analysis","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2406.09246","citing_title":"OpenVLA: An Open-Source Vision-Language-Action Model","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17800","citing_title":"ReFineVLA: Multimodal Reasoning-Aware Generalist Robotic Policies via Teacher-Guided Fine-Tuning","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB","json":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB.json","graph_json":"https://pith.science/api/pith-number/5MQZFVEFUTKRRCVJIB2PNQ56KB/graph.json","events_json":"https://pith.science/api/pith-number/5MQZFVEFUTKRRCVJIB2PNQ56KB/events.json","paper":"https://pith.science/paper/5MQZFVEF"},"agent_actions":{"view_html":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB","download_json":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB.json","view_paper":"https://pith.science/paper/5MQZFVEF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.07280&json=true","fetch_graph":"https://pith.science/api/pith-number/5MQZFVEFUTKRRCVJIB2PNQ56KB/graph.json","fetch_events":"https://pith.science/api/pith-number/5MQZFVEFUTKRRCVJIB2PNQ56KB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB/action/storage_attestation","attest_author":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB/action/author_attestation","sign_citation":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB/action/citation_signature","submit_replication":"https://pith.science/pith/5MQZFVEFUTKRRCVJIB2PNQ56KB/action/replication_record"}},"created_at":"2026-07-05T05:50:34.502021+00:00","updated_at":"2026-07-05T05:50:34.502021+00:00"}