{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3GHBGWNW64DPSV26OGL33QJ72A","short_pith_number":"pith:3GHBGWNW","schema_version":"1.0","canonical_sha256":"d98e1359b6f706f9575e7197bdc13fd0210048a6c932e388bd2b96588dcab816","source":{"kind":"arxiv","id":"2606.21386","version":1},"attestation_state":"computed","paper":{"title":"VLA-FAIL: Efficient Task Failure Detection for Finetuned Vision-Language-Action Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Emiliyan Gospodinov, Enes Ulas Dincer, Florian Seligmann, Gerhard Neumann","submitted_at":"2026-06-19T12:51:21Z","abstract_excerpt":"Vision-language-action models (VLAs) achieve state-of-the-art performance on many robotic manipulation tasks, yet they can still behave unpredictably in out-of-distribution scenarios. Runtime failure detection is therefore essential for the safe real-world deployment of VLAs. However, existing task failure detectors require computationally expensive action sampling, are based on architectural assumptions that limit their applicability to VLAs, or need access to failure rollouts. We propose VLA-FAIL, a lightweight and broadly applicable failure detection framework for VLAs that combines two nov"},"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":"2606.21386","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-19T12:51:21Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"b8691b7894c7b1baacc68c422c6497711b9bd90f2dadba172afba35bd1cc09cb","abstract_canon_sha256":"31580956caca0fd0606d686aac0915ae032624e2eccd2358ceae5e362a1e4f28"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T01:13:08.592220Z","signature_b64":"Z1AoobJZxlp813+ocbjFxakpEjH2xaZ2KHP0DZhnYkagfuyAOyac00+8jtdG7OjEjbtVqxoWgyoUFRNpz2U5Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d98e1359b6f706f9575e7197bdc13fd0210048a6c932e388bd2b96588dcab816","last_reissued_at":"2026-06-23T01:13:08.591730Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T01:13:08.591730Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VLA-FAIL: Efficient Task Failure Detection for Finetuned Vision-Language-Action Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Emiliyan Gospodinov, Enes Ulas Dincer, Florian Seligmann, Gerhard Neumann","submitted_at":"2026-06-19T12:51:21Z","abstract_excerpt":"Vision-language-action models (VLAs) achieve state-of-the-art performance on many robotic manipulation tasks, yet they can still behave unpredictably in out-of-distribution scenarios. Runtime failure detection is therefore essential for the safe real-world deployment of VLAs. However, existing task failure detectors require computationally expensive action sampling, are based on architectural assumptions that limit their applicability to VLAs, or need access to failure rollouts. We propose VLA-FAIL, a lightweight and broadly applicable failure detection framework for VLAs that combines two nov"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.21386","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/2606.21386/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":"2606.21386","created_at":"2026-06-23T01:13:08.591799+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.21386v1","created_at":"2026-06-23T01:13:08.591799+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.21386","created_at":"2026-06-23T01:13:08.591799+00:00"},{"alias_kind":"pith_short_12","alias_value":"3GHBGWNW64DP","created_at":"2026-06-23T01:13:08.591799+00:00"},{"alias_kind":"pith_short_16","alias_value":"3GHBGWNW64DPSV26","created_at":"2026-06-23T01:13:08.591799+00:00"},{"alias_kind":"pith_short_8","alias_value":"3GHBGWNW","created_at":"2026-06-23T01:13:08.591799+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/3GHBGWNW64DPSV26OGL33QJ72A","json":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A.json","graph_json":"https://pith.science/api/pith-number/3GHBGWNW64DPSV26OGL33QJ72A/graph.json","events_json":"https://pith.science/api/pith-number/3GHBGWNW64DPSV26OGL33QJ72A/events.json","paper":"https://pith.science/paper/3GHBGWNW"},"agent_actions":{"view_html":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A","download_json":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A.json","view_paper":"https://pith.science/paper/3GHBGWNW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.21386&json=true","fetch_graph":"https://pith.science/api/pith-number/3GHBGWNW64DPSV26OGL33QJ72A/graph.json","fetch_events":"https://pith.science/api/pith-number/3GHBGWNW64DPSV26OGL33QJ72A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A/action/storage_attestation","attest_author":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A/action/author_attestation","sign_citation":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A/action/citation_signature","submit_replication":"https://pith.science/pith/3GHBGWNW64DPSV26OGL33QJ72A/action/replication_record"}},"created_at":"2026-06-23T01:13:08.591799+00:00","updated_at":"2026-06-23T01:13:08.591799+00:00"}