{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:RYWEI2CX4PSQFA6TAYNZVDUFOS","short_pith_number":"pith:RYWEI2CX","canonical_record":{"source":{"id":"2606.03148","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-02T04:45:02Z","cross_cats_sorted":[],"title_canon_sha256":"f3f42ade27a1391030569a91ba1cbd66617f45a19466490b25a8c639eb5e7a58","abstract_canon_sha256":"f2b2aae46f9790808bafd9ca6cb6cda15134e99985b5f57ff8c8358e3b3319a9"},"schema_version":"1.0"},"canonical_sha256":"8e2c446857e3e50283d3061b9a8e85749dc5324aa48b6398fa5f97889fb6a12a","source":{"kind":"arxiv","id":"2606.03148","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.03148","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"arxiv_version","alias_value":"2606.03148v1","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.03148","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"pith_short_12","alias_value":"RYWEI2CX4PSQ","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"pith_short_16","alias_value":"RYWEI2CX4PSQFA6T","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"pith_short_8","alias_value":"RYWEI2CX","created_at":"2026-06-03T01:05:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:RYWEI2CX4PSQFA6TAYNZVDUFOS","target":"record","payload":{"canonical_record":{"source":{"id":"2606.03148","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-02T04:45:02Z","cross_cats_sorted":[],"title_canon_sha256":"f3f42ade27a1391030569a91ba1cbd66617f45a19466490b25a8c639eb5e7a58","abstract_canon_sha256":"f2b2aae46f9790808bafd9ca6cb6cda15134e99985b5f57ff8c8358e3b3319a9"},"schema_version":"1.0"},"canonical_sha256":"8e2c446857e3e50283d3061b9a8e85749dc5324aa48b6398fa5f97889fb6a12a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T01:05:33.235122Z","signature_b64":"3zIoIxHRhqbBUpbNsuzCv4o+zCerPlpJd6OFBtY3fMNi7ZpF8wFln2FvhI6dSiyN4BXw1oQnZI3XuCX8wb3FDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8e2c446857e3e50283d3061b9a8e85749dc5324aa48b6398fa5f97889fb6a12a","last_reissued_at":"2026-06-03T01:05:33.234725Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T01:05:33.234725Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.03148","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":"2026-06-03T01:05:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"hyw8WZAKjeAehi1Jjqif+BdntgMA9EkcXP6JyThYV0okCGpncmPlw2TFfFNbwOc5iWwJHEN3YgGVmjCfgN1oDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-12T10:12:47.774615Z"},"content_sha256":"dab2ed47a3139ff0b32ef26b1400f6028dee993533189f5a7a81ce8559981b97","schema_version":"1.0","event_id":"sha256:dab2ed47a3139ff0b32ef26b1400f6028dee993533189f5a7a81ce8559981b97"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:RYWEI2CX4PSQFA6TAYNZVDUFOS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"$A^2$: Smaller Self-Supervised ViTs Localize Better than Larger Ones","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Carl Vondrick, Huy Ha, Sreehari Rammohan","submitted_at":"2026-06-02T04:45:02Z","abstract_excerpt":"Robust visual classification often depends on localizing the main foreground objects in an image while ignoring contextual distractors. Surprisingly, we find that the attention maps of smaller self-supervised ViTs localize foreground objects better than those of larger ViTs. However, we still need large ViTs, because they extract richer representations from each patch. To get the best of both worlds, good localization and rich representations, we propose $A^2$, a simple method that leverages this inverse scaling finding by decoupling where to look (a small attention model) from what to extract"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.03148","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/2606.03148/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":"2026-06-03T01:05:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DX0uv+x8uZPAHVNT3RI0FLNVKqglE4XR/MQXRZCiSqysBQv1z2wyX91Q6J3q/TDqXlLikRm61QAe3RF8AYxkBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-12T10:12:47.774990Z"},"content_sha256":"0079a7b7d239a20c7530acf82a54be9a231433969c15d22a3972641e618d77b2","schema_version":"1.0","event_id":"sha256:0079a7b7d239a20c7530acf82a54be9a231433969c15d22a3972641e618d77b2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RYWEI2CX4PSQFA6TAYNZVDUFOS/bundle.json","state_url":"https://pith.science/pith/RYWEI2CX4PSQFA6TAYNZVDUFOS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RYWEI2CX4PSQFA6TAYNZVDUFOS/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":"2026-06-12T10:12:47Z","links":{"resolver":"https://pith.science/pith/RYWEI2CX4PSQFA6TAYNZVDUFOS","bundle":"https://pith.science/pith/RYWEI2CX4PSQFA6TAYNZVDUFOS/bundle.json","state":"https://pith.science/pith/RYWEI2CX4PSQFA6TAYNZVDUFOS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RYWEI2CX4PSQFA6TAYNZVDUFOS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:RYWEI2CX4PSQFA6TAYNZVDUFOS","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":"f2b2aae46f9790808bafd9ca6cb6cda15134e99985b5f57ff8c8358e3b3319a9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-02T04:45:02Z","title_canon_sha256":"f3f42ade27a1391030569a91ba1cbd66617f45a19466490b25a8c639eb5e7a58"},"schema_version":"1.0","source":{"id":"2606.03148","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.03148","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"arxiv_version","alias_value":"2606.03148v1","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.03148","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"pith_short_12","alias_value":"RYWEI2CX4PSQ","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"pith_short_16","alias_value":"RYWEI2CX4PSQFA6T","created_at":"2026-06-03T01:05:33Z"},{"alias_kind":"pith_short_8","alias_value":"RYWEI2CX","created_at":"2026-06-03T01:05:33Z"}],"graph_snapshots":[{"event_id":"sha256:0079a7b7d239a20c7530acf82a54be9a231433969c15d22a3972641e618d77b2","target":"graph","created_at":"2026-06-03T01:05:33Z","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/2606.03148/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Robust visual classification often depends on localizing the main foreground objects in an image while ignoring contextual distractors. Surprisingly, we find that the attention maps of smaller self-supervised ViTs localize foreground objects better than those of larger ViTs. However, we still need large ViTs, because they extract richer representations from each patch. To get the best of both worlds, good localization and rich representations, we propose $A^2$, a simple method that leverages this inverse scaling finding by decoupling where to look (a small attention model) from what to extract","authors_text":"Carl Vondrick, Huy Ha, Sreehari Rammohan","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-02T04:45:02Z","title":"$A^2$: Smaller Self-Supervised ViTs Localize Better than Larger Ones"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.03148","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:dab2ed47a3139ff0b32ef26b1400f6028dee993533189f5a7a81ce8559981b97","target":"record","created_at":"2026-06-03T01:05:33Z","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":"f2b2aae46f9790808bafd9ca6cb6cda15134e99985b5f57ff8c8358e3b3319a9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-02T04:45:02Z","title_canon_sha256":"f3f42ade27a1391030569a91ba1cbd66617f45a19466490b25a8c639eb5e7a58"},"schema_version":"1.0","source":{"id":"2606.03148","kind":"arxiv","version":1}},"canonical_sha256":"8e2c446857e3e50283d3061b9a8e85749dc5324aa48b6398fa5f97889fb6a12a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8e2c446857e3e50283d3061b9a8e85749dc5324aa48b6398fa5f97889fb6a12a","first_computed_at":"2026-06-03T01:05:33.234725Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-03T01:05:33.234725Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"3zIoIxHRhqbBUpbNsuzCv4o+zCerPlpJd6OFBtY3fMNi7ZpF8wFln2FvhI6dSiyN4BXw1oQnZI3XuCX8wb3FDg==","signature_status":"signed_v1","signed_at":"2026-06-03T01:05:33.235122Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.03148","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:dab2ed47a3139ff0b32ef26b1400f6028dee993533189f5a7a81ce8559981b97","sha256:0079a7b7d239a20c7530acf82a54be9a231433969c15d22a3972641e618d77b2"],"state_sha256":"3db189962f9e00a0f265cee5c9578d07ec08cdd7f82965527296ee4d63ecc3fc"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XrFwP8WrQPAdjpb64GLnvvuAXYuSrYPPnleLsKlYpopWE1+FEjXz+RovEYW4Z+WJbn6fe6hMc7pnAb+/nCo4CA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-12T10:12:47.777109Z","bundle_sha256":"9c20fc8d6d2a1e4151f37daf206a726c3709304eaf36a48983e8db678b2bbcc6"}}