{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JETOM6VMXXS4E3IDGNHKDG2YDO","short_pith_number":"pith:JETOM6VM","schema_version":"1.0","canonical_sha256":"4926e67aacbde5c26d03334ea19b581b9f632a920018d25713bce4f6898f263a","source":{"kind":"arxiv","id":"2606.02947","version":1},"attestation_state":"computed","paper":{"title":"BYORn: Bootstrap Your Own Responses to Defend Large Vision-Language Models Against Backdoor Attacks","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Ivan Saboli\\'c, Josip \\v{S}ari\\'c, Marin Or\\v{s}i\\'c, Sven Lon\\v{c}ari\\'c","submitted_at":"2026-06-01T22:58:04Z","abstract_excerpt":"Supervised fine-tuning is the predominant approach for adapting autoregressive vision-language models to downstream tasks. Recent work has shown that this paradigm is highly vulnerable to backdoor attacks, and that existing defenses are ineffective in open-ended generation settings. In response, we propose BYORn, a backdoor-robust fine-tuning framework motivated by the observation that poisoned target responses are often semantically implausible given the corresponding image-text inputs and a pretrained model. BYORn identifies such misaligned responses and dynamically replaces them with altern"},"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":"2606.02947","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-01T22:58:04Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"c65adb0a75cc05c647e76c1ead01b7e5031e10dfe58b341280001bb4f61f3e36","abstract_canon_sha256":"96b39aad9df5d4e8808b2373b55186983edc4a73052432d5722fec14fa140b5c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T01:05:27.245197Z","signature_b64":"XgGkQHOZmwfJso/us9e2EhTavNmPNN34znAXaJcpDrwlZhXYWCk53x2L9GjWwWe+I2+zId/K8Xtytgzq0vajBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4926e67aacbde5c26d03334ea19b581b9f632a920018d25713bce4f6898f263a","last_reissued_at":"2026-06-03T01:05:27.244801Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T01:05:27.244801Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"BYORn: Bootstrap Your Own Responses to Defend Large Vision-Language Models Against Backdoor Attacks","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Ivan Saboli\\'c, Josip \\v{S}ari\\'c, Marin Or\\v{s}i\\'c, Sven Lon\\v{c}ari\\'c","submitted_at":"2026-06-01T22:58:04Z","abstract_excerpt":"Supervised fine-tuning is the predominant approach for adapting autoregressive vision-language models to downstream tasks. Recent work has shown that this paradigm is highly vulnerable to backdoor attacks, and that existing defenses are ineffective in open-ended generation settings. In response, we propose BYORn, a backdoor-robust fine-tuning framework motivated by the observation that poisoned target responses are often semantically implausible given the corresponding image-text inputs and a pretrained model. BYORn identifies such misaligned responses and dynamically replaces them with altern"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02947","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/2606.02947/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":"2606.02947","created_at":"2026-06-03T01:05:27.244872+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02947v1","created_at":"2026-06-03T01:05:27.244872+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02947","created_at":"2026-06-03T01:05:27.244872+00:00"},{"alias_kind":"pith_short_12","alias_value":"JETOM6VMXXS4","created_at":"2026-06-03T01:05:27.244872+00:00"},{"alias_kind":"pith_short_16","alias_value":"JETOM6VMXXS4E3ID","created_at":"2026-06-03T01:05:27.244872+00:00"},{"alias_kind":"pith_short_8","alias_value":"JETOM6VM","created_at":"2026-06-03T01:05:27.244872+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/JETOM6VMXXS4E3IDGNHKDG2YDO","json":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO.json","graph_json":"https://pith.science/api/pith-number/JETOM6VMXXS4E3IDGNHKDG2YDO/graph.json","events_json":"https://pith.science/api/pith-number/JETOM6VMXXS4E3IDGNHKDG2YDO/events.json","paper":"https://pith.science/paper/JETOM6VM"},"agent_actions":{"view_html":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO","download_json":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO.json","view_paper":"https://pith.science/paper/JETOM6VM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02947&json=true","fetch_graph":"https://pith.science/api/pith-number/JETOM6VMXXS4E3IDGNHKDG2YDO/graph.json","fetch_events":"https://pith.science/api/pith-number/JETOM6VMXXS4E3IDGNHKDG2YDO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO/action/storage_attestation","attest_author":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO/action/author_attestation","sign_citation":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO/action/citation_signature","submit_replication":"https://pith.science/pith/JETOM6VMXXS4E3IDGNHKDG2YDO/action/replication_record"}},"created_at":"2026-06-03T01:05:27.244872+00:00","updated_at":"2026-06-03T01:05:27.244872+00:00"}