{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:OPOWPHNAHUAFW4EBE3HLIKGLUJ","short_pith_number":"pith:OPOWPHNA","canonical_record":{"source":{"id":"2605.13322","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:35:07Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"018246d88a61fe17bcde891e1739c69112d2d5f4e748bdc6510af481eee352ac","abstract_canon_sha256":"13de2adc754bafe5a892ae7bf65d2f8ae20a7bcc00474eb6a082691b611a5df6"},"schema_version":"1.0"},"canonical_sha256":"73dd679da03d005b708126ceb428cba241f9591dcb0ef192f72ed6e3901001ce","source":{"kind":"arxiv","id":"2605.13322","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13322","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13322v1","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13322","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"pith_short_12","alias_value":"OPOWPHNAHUAF","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"OPOWPHNAHUAFW4EB","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"OPOWPHNA","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:OPOWPHNAHUAFW4EBE3HLIKGLUJ","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13322","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:35:07Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"018246d88a61fe17bcde891e1739c69112d2d5f4e748bdc6510af481eee352ac","abstract_canon_sha256":"13de2adc754bafe5a892ae7bf65d2f8ae20a7bcc00474eb6a082691b611a5df6"},"schema_version":"1.0"},"canonical_sha256":"73dd679da03d005b708126ceb428cba241f9591dcb0ef192f72ed6e3901001ce","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:48.653115Z","signature_b64":"ySSGmbGPX7J20KUlMqNHmgZ8oHq1rxGphzZZ57C5tt/FN0RRKIuoqdBwGTMafO6m/troc7vkbsGcp+LhziWGAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"73dd679da03d005b708126ceb428cba241f9591dcb0ef192f72ed6e3901001ce","last_reissued_at":"2026-05-18T02:44:48.652722Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:48.652722Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13322","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":"2026-05-18T02:44:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JjXUJjIrqcDT6XRGQhgrlI/4JrTsojZ7VCX6PSgYn3WcwJGtmTwnst/Nf8mmmroZOXomnbymPu+qtuAEdBDABA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-04T19:32:21.792956Z"},"content_sha256":"55f7c735835968f735e33765db43eefc3a7b1e5a9a0b316363168724c154f357","schema_version":"1.0","event_id":"sha256:55f7c735835968f735e33765db43eefc3a7b1e5a9a0b316363168724c154f357"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:OPOWPHNAHUAFW4EBE3HLIKGLUJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"KamonBench: A Grammar-Based Dataset for Evaluating Compositional Factor Recovery in Vision-Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"KamonBench supplies 20,000 grammar-generated crest images whose explicit container, modifier, and motif factors let models be scored directly on compositional recovery rather than caption match alone.","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Richard Sproat, Stefano Peluchetti","submitted_at":"2026-05-13T10:35:07Z","abstract_excerpt":"Kamon (family crests) are an important part of Japanese culture and a natural test case for compositional visual recognition: each crest combines a small number of symbolic choices, but the space of possible descriptions is sparse. We introduce KamonBench, a grammar-based image-to-structure benchmark with 20,000 synthetic composite crests and auxiliary component examples. Each composite crest is paired with a formal kamon description language - \"kamon y\\=ogo\" - description, a segmented Japanese analysis, an English translation, and a non-linguistic program code. Because each synthetic crest is"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"KamonBench therefore provides a controlled testbed for sparse compositional visual recognition and factor recovery in vision-language models.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The grammar rules and synthetic generation process produce images whose factor structure mirrors the compositional challenges present in natural images and real-world visual recognition tasks.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"KamonBench is a grammar-generated synthetic dataset of compositional kamon crests with explicit factor annotations to evaluate factor recovery in vision-language models.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"KamonBench supplies 20,000 grammar-generated crest images whose explicit container, modifier, and motif factors let models be scored directly on compositional recovery rather than caption match alone.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"03e08ce8630d214ae1615d281ae62f130fb6aad4261d7be73377c19a18ab8634"},"source":{"id":"2605.13322","kind":"arxiv","version":1},"verdict":{"id":"d53b069c-3ca8-4f6a-a106-0a32150f6228","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:35:19.971470Z","strongest_claim":"KamonBench therefore provides a controlled testbed for sparse compositional visual recognition and factor recovery in vision-language models.","one_line_summary":"KamonBench is a grammar-generated synthetic dataset of compositional kamon crests with explicit factor annotations to evaluate factor recovery in vision-language models.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The grammar rules and synthetic generation process produce images whose factor structure mirrors the compositional challenges present in natural images and real-world visual recognition tasks.","pith_extraction_headline":"KamonBench supplies 20,000 grammar-generated crest images whose explicit container, modifier, and motif factors let models be scored directly on compositional recovery rather than caption match alone."},"references":{"count":29,"sample":[{"doi":"","year":2017,"title":"Understanding intermediate layers using linear classi- fier probes","work_id":"9d32e99c-c7a8-4a12-8635-d27bff713b31","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1162/coli_a_00422","year":2022,"title":"Probing Classifiers: Promises, Shortcomings, and Advances","work_id":"3eabab74-ac71-4292-86ce-b0469cd4e6cf","ref_index":2,"cited_arxiv_id":"2102.12452","is_internal_anchor":true},{"doi":"","year":1993,"title":"Kadokawa Shoten (角川書店), Tokyo, 1993","work_id":"a978d7c1-72ab-4885-ad0b-94260b0a0831","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale","work_id":"e96730e3-129b-4db6-b981-15ab7932e297","ref_index":4,"cited_arxiv_id":"2010.11929","is_internal_anchor":true},{"doi":"","year":1971,"title":"John Weatherhill, New York, 1971","work_id":"ab8e0fd7-87a3-4a64-a7b3-76fa0a40bf07","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":29,"snapshot_sha256":"4b5904df336149eef7a9411004c35d6460abf6ef773bd5a1f539f40ce6248788","internal_anchors":4},"formal_canon":{"evidence_count":2,"snapshot_sha256":"aa62602e676d9ce1ef388edef9176981f836dd8864815dff05c0db49bdaed63b"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"d53b069c-3ca8-4f6a-a106-0a32150f6228"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GRkltFNzr5w/00ZYwYBh4yCsn1SU1C1e8UMsacFTCfjGd5w5RCzYboD0ri3j3hr668gibeLYKhl1sGPGR6QDCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-04T19:32:21.793592Z"},"content_sha256":"67307e53bfeccf98ca902aac7e9cf43134d740ef5dd44061a25d369b2f609f68","schema_version":"1.0","event_id":"sha256:67307e53bfeccf98ca902aac7e9cf43134d740ef5dd44061a25d369b2f609f68"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/OPOWPHNAHUAFW4EBE3HLIKGLUJ/bundle.json","state_url":"https://pith.science/pith/OPOWPHNAHUAFW4EBE3HLIKGLUJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/OPOWPHNAHUAFW4EBE3HLIKGLUJ/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":"2026-07-04T19:32:21Z","links":{"resolver":"https://pith.science/pith/OPOWPHNAHUAFW4EBE3HLIKGLUJ","bundle":"https://pith.science/pith/OPOWPHNAHUAFW4EBE3HLIKGLUJ/bundle.json","state":"https://pith.science/pith/OPOWPHNAHUAFW4EBE3HLIKGLUJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/OPOWPHNAHUAFW4EBE3HLIKGLUJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:OPOWPHNAHUAFW4EBE3HLIKGLUJ","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":"13de2adc754bafe5a892ae7bf65d2f8ae20a7bcc00474eb6a082691b611a5df6","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:35:07Z","title_canon_sha256":"018246d88a61fe17bcde891e1739c69112d2d5f4e748bdc6510af481eee352ac"},"schema_version":"1.0","source":{"id":"2605.13322","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13322","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13322v1","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13322","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"pith_short_12","alias_value":"OPOWPHNAHUAF","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"OPOWPHNAHUAFW4EB","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"OPOWPHNA","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:67307e53bfeccf98ca902aac7e9cf43134d740ef5dd44061a25d369b2f609f68","target":"graph","created_at":"2026-05-18T02:44:48Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"KamonBench therefore provides a controlled testbed for sparse compositional visual recognition and factor recovery in vision-language models."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The grammar rules and synthetic generation process produce images whose factor structure mirrors the compositional challenges present in natural images and real-world visual recognition tasks."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"KamonBench is a grammar-generated synthetic dataset of compositional kamon crests with explicit factor annotations to evaluate factor recovery in vision-language models."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"KamonBench supplies 20,000 grammar-generated crest images whose explicit container, modifier, and motif factors let models be scored directly on compositional recovery rather than caption match alone."}],"snapshot_sha256":"03e08ce8630d214ae1615d281ae62f130fb6aad4261d7be73377c19a18ab8634"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"aa62602e676d9ce1ef388edef9176981f836dd8864815dff05c0db49bdaed63b"},"paper":{"abstract_excerpt":"Kamon (family crests) are an important part of Japanese culture and a natural test case for compositional visual recognition: each crest combines a small number of symbolic choices, but the space of possible descriptions is sparse. We introduce KamonBench, a grammar-based image-to-structure benchmark with 20,000 synthetic composite crests and auxiliary component examples. Each composite crest is paired with a formal kamon description language - \"kamon y\\=ogo\" - description, a segmented Japanese analysis, an English translation, and a non-linguistic program code. Because each synthetic crest is","authors_text":"Richard Sproat, Stefano Peluchetti","cross_cats":["cs.LG"],"headline":"KamonBench supplies 20,000 grammar-generated crest images whose explicit container, modifier, and motif factors let models be scored directly on compositional recovery rather than caption match alone.","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:35:07Z","title":"KamonBench: A Grammar-Based Dataset for Evaluating Compositional Factor Recovery in Vision-Language Models"},"references":{"count":29,"internal_anchors":4,"resolved_work":29,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Understanding intermediate layers using linear classi- fier probes","work_id":"9d32e99c-c7a8-4a12-8635-d27bff713b31","year":2017},{"cited_arxiv_id":"2102.12452","doi":"10.1162/coli_a_00422","is_internal_anchor":true,"ref_index":2,"title":"Probing Classifiers: Promises, Shortcomings, and Advances","work_id":"3eabab74-ac71-4292-86ce-b0469cd4e6cf","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Kadokawa Shoten (角川書店), Tokyo, 1993","work_id":"a978d7c1-72ab-4885-ad0b-94260b0a0831","year":1993},{"cited_arxiv_id":"2010.11929","doi":"","is_internal_anchor":true,"ref_index":4,"title":"An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale","work_id":"e96730e3-129b-4db6-b981-15ab7932e297","year":2021},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"John Weatherhill, New York, 1971","work_id":"ab8e0fd7-87a3-4a64-a7b3-76fa0a40bf07","year":1971}],"snapshot_sha256":"4b5904df336149eef7a9411004c35d6460abf6ef773bd5a1f539f40ce6248788"},"source":{"id":"2605.13322","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:35:19.971470Z","id":"d53b069c-3ca8-4f6a-a106-0a32150f6228","model_set":{"reader":"grok-4.3"},"one_line_summary":"KamonBench is a grammar-generated synthetic dataset of compositional kamon crests with explicit factor annotations to evaluate factor recovery in vision-language models.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"KamonBench supplies 20,000 grammar-generated crest images whose explicit container, modifier, and motif factors let models be scored directly on compositional recovery rather than caption match alone.","strongest_claim":"KamonBench therefore provides a controlled testbed for sparse compositional visual recognition and factor recovery in vision-language models.","weakest_assumption":"The grammar rules and synthetic generation process produce images whose factor structure mirrors the compositional challenges present in natural images and real-world visual recognition tasks."}},"verdict_id":"d53b069c-3ca8-4f6a-a106-0a32150f6228"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:55f7c735835968f735e33765db43eefc3a7b1e5a9a0b316363168724c154f357","target":"record","created_at":"2026-05-18T02:44:48Z","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":"13de2adc754bafe5a892ae7bf65d2f8ae20a7bcc00474eb6a082691b611a5df6","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:35:07Z","title_canon_sha256":"018246d88a61fe17bcde891e1739c69112d2d5f4e748bdc6510af481eee352ac"},"schema_version":"1.0","source":{"id":"2605.13322","kind":"arxiv","version":1}},"canonical_sha256":"73dd679da03d005b708126ceb428cba241f9591dcb0ef192f72ed6e3901001ce","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"73dd679da03d005b708126ceb428cba241f9591dcb0ef192f72ed6e3901001ce","first_computed_at":"2026-05-18T02:44:48.652722Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:48.652722Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ySSGmbGPX7J20KUlMqNHmgZ8oHq1rxGphzZZ57C5tt/FN0RRKIuoqdBwGTMafO6m/troc7vkbsGcp+LhziWGAg==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:48.653115Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13322","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:55f7c735835968f735e33765db43eefc3a7b1e5a9a0b316363168724c154f357","sha256:67307e53bfeccf98ca902aac7e9cf43134d740ef5dd44061a25d369b2f609f68"],"state_sha256":"0500741b42017c0e7bcbddb2cc384dc2afe01ce666b891a94a9aff368ef7f7a9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Zdp88prx63MKN/yKYR5RqUSeqla9BiRlOus2U6DW6h8rHIkOWyWGsFyjbjJGMZQpago0H2zxbjLRpthSxiBKBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-04T19:32:21.796402Z","bundle_sha256":"971abfc797b56b4d8a3e75d4d5ee87d07919d629406a59ab3b6a941f9b1088a1"}}