{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:SA6ONG4W5GS6WCLXSARBVGL2JU","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"ae953e765b5b113f7cbaa52d5044b344e127084cc7a484b552b4f3d1ce986a7a","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-21T01:02:41Z","title_canon_sha256":"4bcf1b934c4833b105ee8998fcae722fc635c106d143b6f54a17a459cb2cce99"},"schema_version":"1.0","source":{"id":"2605.21854","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.21854","created_at":"2026-05-22T01:04:11Z"},{"alias_kind":"arxiv_version","alias_value":"2605.21854v1","created_at":"2026-05-22T01:04:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21854","created_at":"2026-05-22T01:04:11Z"},{"alias_kind":"pith_short_12","alias_value":"SA6ONG4W5GS6","created_at":"2026-05-22T01:04:11Z"},{"alias_kind":"pith_short_16","alias_value":"SA6ONG4W5GS6WCLX","created_at":"2026-05-22T01:04:11Z"},{"alias_kind":"pith_short_8","alias_value":"SA6ONG4W","created_at":"2026-05-22T01:04:11Z"}],"graph_snapshots":[{"event_id":"sha256:465cef1cb0ea02f0c5564ac08f23526a049564d6d26d45db42cbae255e73c24f","target":"graph","created_at":"2026-05-22T01:04:11Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2605.21854/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Vision-Language-Action (VLA) models have rapidly converged on a small set of architectural patterns: discrete-token autoregression (e.g. OpenVLA) and continuous-action flow-matching (e.g. pi-0.5). Yet preference alignment via Direct Preference Optimisation (DPO) -- the de-facto post-training step in language models -- has been studied almost exclusively on autoregressive VLAs. We present CrossVLA, an empirical study of cross-paradigm VLA post-training. Three contributions: (i) a surrogate flow-matching log-probability estimator that lets DPO operate on continuous-action backbones without proba","authors_text":"Zhi Liu","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-21T01:02:41Z","title":"CrossVLA: Cross-Paradigm Post-Training and Inference Optimization for Vision-Language-Action Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21854","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6280820c9c41ef0a3fc0918255231690cd560358277bdb398ca4f3721f3feaac","target":"record","created_at":"2026-05-22T01:04:11Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"ae953e765b5b113f7cbaa52d5044b344e127084cc7a484b552b4f3d1ce986a7a","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-21T01:02:41Z","title_canon_sha256":"4bcf1b934c4833b105ee8998fcae722fc635c106d143b6f54a17a459cb2cce99"},"schema_version":"1.0","source":{"id":"2605.21854","kind":"arxiv","version":1}},"canonical_sha256":"903ce69b96e9a5eb097790221a997a4d0fb9eead318a23be162976aec14ca7b5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"903ce69b96e9a5eb097790221a997a4d0fb9eead318a23be162976aec14ca7b5","first_computed_at":"2026-05-22T01:04:11.469670Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-22T01:04:11.469670Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HtNsDqZKzeU0SCZaTKOSadsH+mdsIChidT/ZtJbbv+gW2YGxtmePaoe1BhOAHs5naf7U/GUcXHPxxUoYj5DrCA==","signature_status":"signed_v1","signed_at":"2026-05-22T01:04:11.470367Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.21854","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6280820c9c41ef0a3fc0918255231690cd560358277bdb398ca4f3721f3feaac","sha256:465cef1cb0ea02f0c5564ac08f23526a049564d6d26d45db42cbae255e73c24f"],"state_sha256":"c3d0311c372f1ceccdc04f916a00f374c743c3de5099a705d6d509289fb1f3df"}