{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QGGVUFRJRT5MTVW5BLZBJOIYXB","short_pith_number":"pith:QGGVUFRJ","schema_version":"1.0","canonical_sha256":"818d5a16298cfac9d6dd0af214b918b84df5bdf69174927c7324f5fa181715cf","source":{"kind":"arxiv","id":"2608.04246","version":1},"attestation_state":"computed","paper":{"title":"SAFECAST: Robust Failure Detection for VLA Policies with Contrast-Set Training and Calibration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Abrar Anwar, Aditeya Prajapati, Harshitha Rajaprakash, Jesse Thomason, Rong Xue","submitted_at":"2026-08-04T21:57:03Z","abstract_excerpt":"Vision-language-action policies often fail under deployment-time distribution shifts such as clutter, distractor objects, lighting changes, novel objects, altered initial states, and reworded instructions. Hidden-state-based risk probes combined with functional conformal prediction can detect rollout failures, but their reliability depends on calibration data matching deployment conditions. We introduce SAFECAST, which leverages contrast set perturbations to improve hidden-state probe training and calibration for deployment time shift. SAFECAST statistically significantly improves failure dete"},"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":"2608.04246","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-08-04T21:57:03Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"4f708593d32bd9829bead7fec98f3a307433dab0c6ed2e41f3a70bb05d75b903","abstract_canon_sha256":"5465d9db9c796601af5618bf06116be35db126792b82a40006a7180e2f0ece29"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-06T01:31:44.552046Z","signature_b64":"q++X1EYrNV9j/2IMdSZdgSG1OOfcazxeFEeES9hIQFXfSur1TO/m0iKQKw/ZnhEUtVZpzv71epuhexBRpPqtAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"818d5a16298cfac9d6dd0af214b918b84df5bdf69174927c7324f5fa181715cf","last_reissued_at":"2026-08-06T01:31:44.550539Z","signature_status":"signed_v1","first_computed_at":"2026-08-06T01:31:44.550539Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SAFECAST: Robust Failure Detection for VLA Policies with Contrast-Set Training and Calibration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Abrar Anwar, Aditeya Prajapati, Harshitha Rajaprakash, Jesse Thomason, Rong Xue","submitted_at":"2026-08-04T21:57:03Z","abstract_excerpt":"Vision-language-action policies often fail under deployment-time distribution shifts such as clutter, distractor objects, lighting changes, novel objects, altered initial states, and reworded instructions. Hidden-state-based risk probes combined with functional conformal prediction can detect rollout failures, but their reliability depends on calibration data matching deployment conditions. We introduce SAFECAST, which leverages contrast set perturbations to improve hidden-state probe training and calibration for deployment time shift. SAFECAST statistically significantly improves failure dete"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.04246","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/2608.04246/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":"2608.04246","created_at":"2026-08-06T01:31:44.552384+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.04246v1","created_at":"2026-08-06T01:31:44.552384+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.04246","created_at":"2026-08-06T01:31:44.552384+00:00"},{"alias_kind":"pith_short_12","alias_value":"QGGVUFRJRT5M","created_at":"2026-08-06T01:31:44.552384+00:00"},{"alias_kind":"pith_short_16","alias_value":"QGGVUFRJRT5MTVW5","created_at":"2026-08-06T01:31:44.552384+00:00"},{"alias_kind":"pith_short_8","alias_value":"QGGVUFRJ","created_at":"2026-08-06T01:31:44.552384+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB","json":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB.json","graph_json":"https://pith.science/api/pith-number/QGGVUFRJRT5MTVW5BLZBJOIYXB/graph.json","events_json":"https://pith.science/api/pith-number/QGGVUFRJRT5MTVW5BLZBJOIYXB/events.json","paper":"https://pith.science/paper/QGGVUFRJ"},"agent_actions":{"view_html":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB","download_json":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB.json","view_paper":"https://pith.science/paper/QGGVUFRJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.04246&json=true","fetch_graph":"https://pith.science/api/pith-number/QGGVUFRJRT5MTVW5BLZBJOIYXB/graph.json","fetch_events":"https://pith.science/api/pith-number/QGGVUFRJRT5MTVW5BLZBJOIYXB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB/action/storage_attestation","attest_author":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB/action/author_attestation","sign_citation":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB/action/citation_signature","submit_replication":"https://pith.science/pith/QGGVUFRJRT5MTVW5BLZBJOIYXB/action/replication_record"}},"created_at":"2026-08-06T01:31:44.552384+00:00","updated_at":"2026-08-06T01:31:44.552384+00:00"}