{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TH553YQP6JDUQLZYGMK2FKBAGR","short_pith_number":"pith:TH553YQP","schema_version":"1.0","canonical_sha256":"99fbdde20ff247482f383315a2a820347656ceb80f3128138278c11952f24b05","source":{"kind":"arxiv","id":"2606.27755","version":1},"attestation_state":"computed","paper":{"title":"Drop-Then-Recovery: How Redundant Are Vision-Language-Action Models?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Ang Li, Gaowen Liu, Guoheng Sun, Kaixi Feng, Ramana Rao Kompella, Shwai He, Wanghao Ye, Xiaochuan Gong, Yexiao He, Zheyu Shen, Ziyao Wang","submitted_at":"2026-06-26T06:22:17Z","abstract_excerpt":"Vision-Language-Action (VLA) models enable instruction-driven robotic manipulation, but they inherit oversized language backbones from pretrained VLMs whose capacity far exceeds what is needed for short robotic instructions. This raises a basic question: how much of a VLA model is actually necessary for closed-loop control? In this work, we study architectural redundancy in VLA models by using transformer block removal as a controlled intervention. We introduce \\textbf{Drop-Then-Recovery (DTR)}, an analysis protocol that removes selected blocks from a pretrained VLA model and then fine-tunes t"},"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":"2606.27755","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-06-26T06:22:17Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"651970af1eca4684a7787cf48cf36b926812de6a7195e55f85308c3a21b6645a","abstract_canon_sha256":"dda8fcefaa0cf29c2471ac6e95036b9b557f45ab1a4b64270114fa7cfed912fa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-29T01:14:47.544663Z","signature_b64":"g2DGVT8H3dGZZyBm2GltFc8+2Jlpya/E3yMiZLOUtOpi4VNLSWfAXPm1x+Xm0lp54MsvDHUpg2az5lma0NIzDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"99fbdde20ff247482f383315a2a820347656ceb80f3128138278c11952f24b05","last_reissued_at":"2026-06-29T01:14:47.544272Z","signature_status":"signed_v1","first_computed_at":"2026-06-29T01:14:47.544272Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Drop-Then-Recovery: How Redundant Are Vision-Language-Action Models?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Ang Li, Gaowen Liu, Guoheng Sun, Kaixi Feng, Ramana Rao Kompella, Shwai He, Wanghao Ye, Xiaochuan Gong, Yexiao He, Zheyu Shen, Ziyao Wang","submitted_at":"2026-06-26T06:22:17Z","abstract_excerpt":"Vision-Language-Action (VLA) models enable instruction-driven robotic manipulation, but they inherit oversized language backbones from pretrained VLMs whose capacity far exceeds what is needed for short robotic instructions. This raises a basic question: how much of a VLA model is actually necessary for closed-loop control? In this work, we study architectural redundancy in VLA models by using transformer block removal as a controlled intervention. We introduce \\textbf{Drop-Then-Recovery (DTR)}, an analysis protocol that removes selected blocks from a pretrained VLA model and then fine-tunes t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.27755","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/2606.27755/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":"2606.27755","created_at":"2026-06-29T01:14:47.544327+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.27755v1","created_at":"2026-06-29T01:14:47.544327+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.27755","created_at":"2026-06-29T01:14:47.544327+00:00"},{"alias_kind":"pith_short_12","alias_value":"TH553YQP6JDU","created_at":"2026-06-29T01:14:47.544327+00:00"},{"alias_kind":"pith_short_16","alias_value":"TH553YQP6JDUQLZY","created_at":"2026-06-29T01:14:47.544327+00:00"},{"alias_kind":"pith_short_8","alias_value":"TH553YQP","created_at":"2026-06-29T01:14:47.544327+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/TH553YQP6JDUQLZYGMK2FKBAGR","json":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR.json","graph_json":"https://pith.science/api/pith-number/TH553YQP6JDUQLZYGMK2FKBAGR/graph.json","events_json":"https://pith.science/api/pith-number/TH553YQP6JDUQLZYGMK2FKBAGR/events.json","paper":"https://pith.science/paper/TH553YQP"},"agent_actions":{"view_html":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR","download_json":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR.json","view_paper":"https://pith.science/paper/TH553YQP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.27755&json=true","fetch_graph":"https://pith.science/api/pith-number/TH553YQP6JDUQLZYGMK2FKBAGR/graph.json","fetch_events":"https://pith.science/api/pith-number/TH553YQP6JDUQLZYGMK2FKBAGR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR/action/storage_attestation","attest_author":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR/action/author_attestation","sign_citation":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR/action/citation_signature","submit_replication":"https://pith.science/pith/TH553YQP6JDUQLZYGMK2FKBAGR/action/replication_record"}},"created_at":"2026-06-29T01:14:47.544327+00:00","updated_at":"2026-06-29T01:14:47.544327+00:00"}