{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CMWVBVGA24G4HB2EP3HMR4N7IK","short_pith_number":"pith:CMWVBVGA","schema_version":"1.0","canonical_sha256":"132d50d4c0d70dc387447ecec8f1bf428d356f20a03603fe4ece418383927687","source":{"kind":"arxiv","id":"2605.30117","version":1},"attestation_state":"computed","paper":{"title":"VLA-Trace: Diagnosing Vision-Language-Action Models through Representation and Behavior Tracing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Han Dong, Haoyuan Shi, Haozhe Shan, Jiayu Hu, Jinpeng Lu, Qinfan Zhang, Xiancong Ren, Xiaozhu Ju, Yinda Chen, Yingji Zhang, Yi Zhang, Yong Dai","submitted_at":"2026-05-28T15:50:56Z","abstract_excerpt":"Understanding how Vision-Language-Action (VLA) models transform multimodal knowledge into embodied control remains an open challenge. We present VLA-Trace, a progressive diagnostic framework that analyzes VLA models through a unified evidence chain from representation dynamics to causal control attribution and behavioral manifestation. It specifically combines cross-modal and checkpoint-drift centered kernel alignment (CKA) to trace representation evolution, attention knockout interventions to identify modality-specific control pathways, and rollout-level behavioral probes to examine grounding"},"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":"2605.30117","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-28T15:50:56Z","cross_cats_sorted":[],"title_canon_sha256":"eb21bfec5723b36c0c3576b2e2c5894896d81cb8a2f3369752400867d1802e7b","abstract_canon_sha256":"3bd113d5ea59cfd3a061785099bb09d474d5bf558c9d073d7fdee48f8e229297"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T02:06:10.271504Z","signature_b64":"oqXkW5cOGyzXOEmYisZg6OlbNXhGuG/9ZkIenphpnwXtvT9wk68IOzlkuNkyVILYE1RZR4lCMsEDQE4aIk/XDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"132d50d4c0d70dc387447ecec8f1bf428d356f20a03603fe4ece418383927687","last_reissued_at":"2026-05-29T02:06:10.271143Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T02:06:10.271143Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VLA-Trace: Diagnosing Vision-Language-Action Models through Representation and Behavior Tracing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Han Dong, Haoyuan Shi, Haozhe Shan, Jiayu Hu, Jinpeng Lu, Qinfan Zhang, Xiancong Ren, Xiaozhu Ju, Yinda Chen, Yingji Zhang, Yi Zhang, Yong Dai","submitted_at":"2026-05-28T15:50:56Z","abstract_excerpt":"Understanding how Vision-Language-Action (VLA) models transform multimodal knowledge into embodied control remains an open challenge. We present VLA-Trace, a progressive diagnostic framework that analyzes VLA models through a unified evidence chain from representation dynamics to causal control attribution and behavioral manifestation. It specifically combines cross-modal and checkpoint-drift centered kernel alignment (CKA) to trace representation evolution, attention knockout interventions to identify modality-specific control pathways, and rollout-level behavioral probes to examine grounding"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30117","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/2605.30117/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":"2605.30117","created_at":"2026-05-29T02:06:10.271205+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.30117v1","created_at":"2026-05-29T02:06:10.271205+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30117","created_at":"2026-05-29T02:06:10.271205+00:00"},{"alias_kind":"pith_short_12","alias_value":"CMWVBVGA24G4","created_at":"2026-05-29T02:06:10.271205+00:00"},{"alias_kind":"pith_short_16","alias_value":"CMWVBVGA24G4HB2E","created_at":"2026-05-29T02:06:10.271205+00:00"},{"alias_kind":"pith_short_8","alias_value":"CMWVBVGA","created_at":"2026-05-29T02:06:10.271205+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/CMWVBVGA24G4HB2EP3HMR4N7IK","json":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK.json","graph_json":"https://pith.science/api/pith-number/CMWVBVGA24G4HB2EP3HMR4N7IK/graph.json","events_json":"https://pith.science/api/pith-number/CMWVBVGA24G4HB2EP3HMR4N7IK/events.json","paper":"https://pith.science/paper/CMWVBVGA"},"agent_actions":{"view_html":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK","download_json":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK.json","view_paper":"https://pith.science/paper/CMWVBVGA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.30117&json=true","fetch_graph":"https://pith.science/api/pith-number/CMWVBVGA24G4HB2EP3HMR4N7IK/graph.json","fetch_events":"https://pith.science/api/pith-number/CMWVBVGA24G4HB2EP3HMR4N7IK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK/action/storage_attestation","attest_author":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK/action/author_attestation","sign_citation":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK/action/citation_signature","submit_replication":"https://pith.science/pith/CMWVBVGA24G4HB2EP3HMR4N7IK/action/replication_record"}},"created_at":"2026-05-29T02:06:10.271205+00:00","updated_at":"2026-05-29T02:06:10.271205+00:00"}