{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:I76H4M62PBHENB7DG3MWDTIW6C","short_pith_number":"pith:I76H4M62","canonical_record":{"source":{"id":"2505.18503","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-24T04:45:45Z","cross_cats_sorted":[],"title_canon_sha256":"66408954edf9f26bde0f52f490a53c1a006e665714e8312ecf54fbc22d848ea3","abstract_canon_sha256":"2e69f5f868a60c404b024c45c4a1c86386632cc0882a8c9ec3ad14aafe85c1fe"},"schema_version":"1.0"},"canonical_sha256":"47fc7e33da784e4687e336d961cd16f0aabcab1bec90adc1ecab4a7ff9b1ae80","source":{"kind":"arxiv","id":"2505.18503","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.18503","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"arxiv_version","alias_value":"2505.18503v1","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18503","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"pith_short_12","alias_value":"I76H4M62PBHE","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"pith_short_16","alias_value":"I76H4M62PBHENB7D","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"pith_short_8","alias_value":"I76H4M62","created_at":"2026-07-05T11:08:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:I76H4M62PBHENB7DG3MWDTIW6C","target":"record","payload":{"canonical_record":{"source":{"id":"2505.18503","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-24T04:45:45Z","cross_cats_sorted":[],"title_canon_sha256":"66408954edf9f26bde0f52f490a53c1a006e665714e8312ecf54fbc22d848ea3","abstract_canon_sha256":"2e69f5f868a60c404b024c45c4a1c86386632cc0882a8c9ec3ad14aafe85c1fe"},"schema_version":"1.0"},"canonical_sha256":"47fc7e33da784e4687e336d961cd16f0aabcab1bec90adc1ecab4a7ff9b1ae80","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:08:47.389155Z","signature_b64":"oqGSaRkVRw6AFiHPGlKqBqpGYx3hutQjamoqVAcnvhFxF5nLr+a6YJrvjC3QCM3bQ4K+cn6xca4xwyeXWMDLCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"47fc7e33da784e4687e336d961cd16f0aabcab1bec90adc1ecab4a7ff9b1ae80","last_reissued_at":"2026-07-05T11:08:47.388639Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:08:47.388639Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.18503","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:08:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mWzFA7yCO3VoTfn+93bFIU2HPSwIafYgiPWsE2XY4Me4tfQkLpmn1SlcvXs6E3i0lnFvZqalA0cIWegBrHIMDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T16:34:03.170330Z"},"content_sha256":"e452bc0dd90e6ed386915fd7c3e8ea6e22358b0ea9ee05baa9b444abe8621f09","schema_version":"1.0","event_id":"sha256:e452bc0dd90e6ed386915fd7c3e8ea6e22358b0ea9ee05baa9b444abe8621f09"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:I76H4M62PBHENB7DG3MWDTIW6C","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Focus on What Matters: Enhancing Medical Vision-Language Models with Automatic Attention Alignment Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Alex James Boyd, Aofei Chang, Cao Xiao, Fenglong Ma, Le Huang, Parminder Bhatia, Taha Kass-hout","submitted_at":"2025-05-24T04:45:45Z","abstract_excerpt":"Medical Large Vision-Language Models (Med-LVLMs) often exhibit suboptimal attention distribution on visual inputs, leading to hallucinated or inaccurate outputs. Existing mitigation methods primarily rely on inference-time interventions, which are limited in attention adaptation or require additional supervision. To address this, we propose A$^3$Tune, a novel fine-tuning framework for Automatic Attention Alignment Tuning. A$^3$Tune leverages zero-shot weak labels from SAM, refines them into prompt-aware labels using BioMedCLIP, and then selectively modifies visually-critical attention heads to"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18503","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/2505.18503/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:08:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uqLBkk6jbS8Jt5eE1W8ZNs+vR1zhaCY/z0Th4PmpDl835LmzLBJyM3L/TXuos63bBVv0UMBcIhIQiAMyAi49BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T16:34:03.171234Z"},"content_sha256":"5cc624c1c32300beff69ef1018ea878385e6d47660a11368979d04d43ca58dfb","schema_version":"1.0","event_id":"sha256:5cc624c1c32300beff69ef1018ea878385e6d47660a11368979d04d43ca58dfb"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/I76H4M62PBHENB7DG3MWDTIW6C/bundle.json","state_url":"https://pith.science/pith/I76H4M62PBHENB7DG3MWDTIW6C/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/I76H4M62PBHENB7DG3MWDTIW6C/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-03T16:34:03Z","links":{"resolver":"https://pith.science/pith/I76H4M62PBHENB7DG3MWDTIW6C","bundle":"https://pith.science/pith/I76H4M62PBHENB7DG3MWDTIW6C/bundle.json","state":"https://pith.science/pith/I76H4M62PBHENB7DG3MWDTIW6C/state.json","well_known_bundle":"https://pith.science/.well-known/pith/I76H4M62PBHENB7DG3MWDTIW6C/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:I76H4M62PBHENB7DG3MWDTIW6C","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":"2e69f5f868a60c404b024c45c4a1c86386632cc0882a8c9ec3ad14aafe85c1fe","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-24T04:45:45Z","title_canon_sha256":"66408954edf9f26bde0f52f490a53c1a006e665714e8312ecf54fbc22d848ea3"},"schema_version":"1.0","source":{"id":"2505.18503","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.18503","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"arxiv_version","alias_value":"2505.18503v1","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18503","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"pith_short_12","alias_value":"I76H4M62PBHE","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"pith_short_16","alias_value":"I76H4M62PBHENB7D","created_at":"2026-07-05T11:08:47Z"},{"alias_kind":"pith_short_8","alias_value":"I76H4M62","created_at":"2026-07-05T11:08:47Z"}],"graph_snapshots":[{"event_id":"sha256:5cc624c1c32300beff69ef1018ea878385e6d47660a11368979d04d43ca58dfb","target":"graph","created_at":"2026-07-05T11:08:47Z","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/2505.18503/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Medical Large Vision-Language Models (Med-LVLMs) often exhibit suboptimal attention distribution on visual inputs, leading to hallucinated or inaccurate outputs. Existing mitigation methods primarily rely on inference-time interventions, which are limited in attention adaptation or require additional supervision. To address this, we propose A$^3$Tune, a novel fine-tuning framework for Automatic Attention Alignment Tuning. A$^3$Tune leverages zero-shot weak labels from SAM, refines them into prompt-aware labels using BioMedCLIP, and then selectively modifies visually-critical attention heads to","authors_text":"Alex James Boyd, Aofei Chang, Cao Xiao, Fenglong Ma, Le Huang, Parminder Bhatia, Taha Kass-hout","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-24T04:45:45Z","title":"Focus on What Matters: Enhancing Medical Vision-Language Models with Automatic Attention Alignment Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18503","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:e452bc0dd90e6ed386915fd7c3e8ea6e22358b0ea9ee05baa9b444abe8621f09","target":"record","created_at":"2026-07-05T11:08:47Z","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":"2e69f5f868a60c404b024c45c4a1c86386632cc0882a8c9ec3ad14aafe85c1fe","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-24T04:45:45Z","title_canon_sha256":"66408954edf9f26bde0f52f490a53c1a006e665714e8312ecf54fbc22d848ea3"},"schema_version":"1.0","source":{"id":"2505.18503","kind":"arxiv","version":1}},"canonical_sha256":"47fc7e33da784e4687e336d961cd16f0aabcab1bec90adc1ecab4a7ff9b1ae80","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"47fc7e33da784e4687e336d961cd16f0aabcab1bec90adc1ecab4a7ff9b1ae80","first_computed_at":"2026-07-05T11:08:47.388639Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:08:47.388639Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"oqGSaRkVRw6AFiHPGlKqBqpGYx3hutQjamoqVAcnvhFxF5nLr+a6YJrvjC3QCM3bQ4K+cn6xca4xwyeXWMDLCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:08:47.389155Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.18503","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e452bc0dd90e6ed386915fd7c3e8ea6e22358b0ea9ee05baa9b444abe8621f09","sha256:5cc624c1c32300beff69ef1018ea878385e6d47660a11368979d04d43ca58dfb"],"state_sha256":"35b2a097a133a2fb6aeef9b03506d3ba912747edafa101d089705a102c07d2c6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ownhc+cQiKTVMjWJjt1AwcmCknpFUbdGM5J6fZCOOKo/YGI2GOFjTClcV6fHQSEnE4Uxq3YG8xdMEIZHqgM+DA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T16:34:03.176443Z","bundle_sha256":"476adf7d47944aefe2697e6759c344f64dff39ec629489386e3e60117b4902a2"}}