{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:KH7FCWR5MKMOSPF4PSNZL7BBIO","short_pith_number":"pith:KH7FCWR5","canonical_record":{"source":{"id":"2503.06287","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-08T17:24:42Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a38cb8808b7f7f462b1ebcf4a6637d7bdeaff40717f0f640b33db387b4528a4d","abstract_canon_sha256":"14ff29606c38179cb21ae1e114882e0d755210d998899c1686336da3db0f22b1"},"schema_version":"1.0"},"canonical_sha256":"51fe515a3d6298e93cbc7c9b95fc2143869308f69df740428d9c69f221db0ffb","source":{"kind":"arxiv","id":"2503.06287","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06287","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06287v1","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06287","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"pith_short_12","alias_value":"KH7FCWR5MKMO","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"pith_short_16","alias_value":"KH7FCWR5MKMOSPF4","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"pith_short_8","alias_value":"KH7FCWR5","created_at":"2026-07-05T10:27:29Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:KH7FCWR5MKMOSPF4PSNZL7BBIO","target":"record","payload":{"canonical_record":{"source":{"id":"2503.06287","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-08T17:24:42Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a38cb8808b7f7f462b1ebcf4a6637d7bdeaff40717f0f640b33db387b4528a4d","abstract_canon_sha256":"14ff29606c38179cb21ae1e114882e0d755210d998899c1686336da3db0f22b1"},"schema_version":"1.0"},"canonical_sha256":"51fe515a3d6298e93cbc7c9b95fc2143869308f69df740428d9c69f221db0ffb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:29.840087Z","signature_b64":"+x/otuNGenHbMb94fsCt6ZgFeINskJJSmOU8bgM79+3UDwC83ysy+11+WaCA3CVRcJSXT+Ld7KdNk95AL3+QBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"51fe515a3d6298e93cbc7c9b95fc2143869308f69df740428d9c69f221db0ffb","last_reissued_at":"2026-07-05T10:27:29.839467Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:29.839467Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.06287","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":"2026-07-05T10:27:29Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XHFMxYikP9UIcfY7fTsLXokMS2jp+Bu2MftFOyD6+d7tg9GPWk9szcoSiL7uK4uEA63faEAYnn2C73Omju/oDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T02:41:54.056208Z"},"content_sha256":"b1ea81d5047d17bad640ee22d76da8ac34a40451f5affa28d67820b42e1f0384","schema_version":"1.0","event_id":"sha256:b1ea81d5047d17bad640ee22d76da8ac34a40451f5affa28d67820b42e1f0384"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:KH7FCWR5MKMOSPF4PSNZL7BBIO","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Your Large Vision-Language Model Only Needs A Few Attention Heads For Visual Grounding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jinyeong Kim, Junhyeok Kim, Seil Kang, Seong Jae Hwang","submitted_at":"2025-03-08T17:24:42Z","abstract_excerpt":"Visual grounding seeks to localize the image region corresponding to a free-form text description. Recently, the strong multimodal capabilities of Large Vision-Language Models (LVLMs) have driven substantial improvements in visual grounding, though they inevitably require fine-tuning and additional model components to explicitly generate bounding boxes or segmentation masks. However, we discover that a few attention heads in frozen LVLMs demonstrate strong visual grounding capabilities. We refer to these heads, which consistently capture object locations related to text semantics, as localizat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06287","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/2503.06287/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":"2026-07-05T10:27:29Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YLLTh53lFvOcH+cv+eLNa+LPNt5W6W2SazoDdpmtNMwUZXNgLa3J+jiwLtRmxNAp607UToJaRXs9urnawoJHCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T02:41:54.056848Z"},"content_sha256":"88ddc53f1d3a3d1a026272070ab88c5c48567ead36808c3d3ab16db2e4df0990","schema_version":"1.0","event_id":"sha256:88ddc53f1d3a3d1a026272070ab88c5c48567ead36808c3d3ab16db2e4df0990"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KH7FCWR5MKMOSPF4PSNZL7BBIO/bundle.json","state_url":"https://pith.science/pith/KH7FCWR5MKMOSPF4PSNZL7BBIO/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KH7FCWR5MKMOSPF4PSNZL7BBIO/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":"2026-07-07T02:41:54Z","links":{"resolver":"https://pith.science/pith/KH7FCWR5MKMOSPF4PSNZL7BBIO","bundle":"https://pith.science/pith/KH7FCWR5MKMOSPF4PSNZL7BBIO/bundle.json","state":"https://pith.science/pith/KH7FCWR5MKMOSPF4PSNZL7BBIO/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KH7FCWR5MKMOSPF4PSNZL7BBIO/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:KH7FCWR5MKMOSPF4PSNZL7BBIO","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":"14ff29606c38179cb21ae1e114882e0d755210d998899c1686336da3db0f22b1","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-08T17:24:42Z","title_canon_sha256":"a38cb8808b7f7f462b1ebcf4a6637d7bdeaff40717f0f640b33db387b4528a4d"},"schema_version":"1.0","source":{"id":"2503.06287","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06287","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06287v1","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06287","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"pith_short_12","alias_value":"KH7FCWR5MKMO","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"pith_short_16","alias_value":"KH7FCWR5MKMOSPF4","created_at":"2026-07-05T10:27:29Z"},{"alias_kind":"pith_short_8","alias_value":"KH7FCWR5","created_at":"2026-07-05T10:27:29Z"}],"graph_snapshots":[{"event_id":"sha256:88ddc53f1d3a3d1a026272070ab88c5c48567ead36808c3d3ab16db2e4df0990","target":"graph","created_at":"2026-07-05T10:27:29Z","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/2503.06287/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Visual grounding seeks to localize the image region corresponding to a free-form text description. Recently, the strong multimodal capabilities of Large Vision-Language Models (LVLMs) have driven substantial improvements in visual grounding, though they inevitably require fine-tuning and additional model components to explicitly generate bounding boxes or segmentation masks. However, we discover that a few attention heads in frozen LVLMs demonstrate strong visual grounding capabilities. We refer to these heads, which consistently capture object locations related to text semantics, as localizat","authors_text":"Jinyeong Kim, Junhyeok Kim, Seil Kang, Seong Jae Hwang","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-08T17:24:42Z","title":"Your Large Vision-Language Model Only Needs A Few Attention Heads For Visual Grounding"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06287","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:b1ea81d5047d17bad640ee22d76da8ac34a40451f5affa28d67820b42e1f0384","target":"record","created_at":"2026-07-05T10:27:29Z","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":"14ff29606c38179cb21ae1e114882e0d755210d998899c1686336da3db0f22b1","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-08T17:24:42Z","title_canon_sha256":"a38cb8808b7f7f462b1ebcf4a6637d7bdeaff40717f0f640b33db387b4528a4d"},"schema_version":"1.0","source":{"id":"2503.06287","kind":"arxiv","version":1}},"canonical_sha256":"51fe515a3d6298e93cbc7c9b95fc2143869308f69df740428d9c69f221db0ffb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"51fe515a3d6298e93cbc7c9b95fc2143869308f69df740428d9c69f221db0ffb","first_computed_at":"2026-07-05T10:27:29.839467Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:27:29.839467Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"+x/otuNGenHbMb94fsCt6ZgFeINskJJSmOU8bgM79+3UDwC83ysy+11+WaCA3CVRcJSXT+Ld7KdNk95AL3+QBA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:27:29.840087Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.06287","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b1ea81d5047d17bad640ee22d76da8ac34a40451f5affa28d67820b42e1f0384","sha256:88ddc53f1d3a3d1a026272070ab88c5c48567ead36808c3d3ab16db2e4df0990"],"state_sha256":"4dfcfb5e95ddc168ccf63f8b905d39e20219e20422493f67e2ae6d9cf774917f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rJlxMwefAO52WZvTLVCNjbaWfer6376pWI9n4TNiioeBlegzFU8aVEqvXlDpdQv/gxTTgFACif9tkCsbNPevAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T02:41:54.059419Z","bundle_sha256":"ee57da7bd4b82307bdde591bd8317ad061657fa43d94dd429dd2133ec83c61af"}}