{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:GJ4YTDA7XX7Y6QSJ7DFIMBQS2O","short_pith_number":"pith:GJ4YTDA7","canonical_record":{"source":{"id":"2601.04946","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-01-08T13:49:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4e5bd205d625ef46a2157ee876c341c44ba59b4adeb09d55b77a67b83191707f","abstract_canon_sha256":"e1a3df6b302157012e13eb6c3a01aca6eaa5d69c5f7f2c6cdb843e3903469d32"},"schema_version":"1.0"},"canonical_sha256":"3279898c1fbdff8f4249f8ca860612d38b42e5b180faa061004acb883b134554","source":{"kind":"arxiv","id":"2601.04946","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.04946","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2601.04946v3","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.04946","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"GJ4YTDA7XX7Y","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"GJ4YTDA7XX7Y6QSJ","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"GJ4YTDA7","created_at":"2026-06-02T03:04:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:GJ4YTDA7XX7Y6QSJ7DFIMBQS2O","target":"record","payload":{"canonical_record":{"source":{"id":"2601.04946","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-01-08T13:49:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4e5bd205d625ef46a2157ee876c341c44ba59b4adeb09d55b77a67b83191707f","abstract_canon_sha256":"e1a3df6b302157012e13eb6c3a01aca6eaa5d69c5f7f2c6cdb843e3903469d32"},"schema_version":"1.0"},"canonical_sha256":"3279898c1fbdff8f4249f8ca860612d38b42e5b180faa061004acb883b134554","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T03:04:36.872555Z","signature_b64":"LmKBfz5ZS0y78OCHqMaudlmF4qkeepBmKSCekmQ63XlzK/HPTNaEYhP22UL4AyBW6nvfYgC6aNt7r/jsu+ZBAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3279898c1fbdff8f4249f8ca860612d38b42e5b180faa061004acb883b134554","last_reissued_at":"2026-06-02T03:04:36.872044Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T03:04:36.872044Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2601.04946","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T03:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Wjws4kg/IDRlN+SqcZx1y0WOurl5tLpJlAfQoioJyrHRfTO+Ag91OQKL3Zp/AR9SXM83/P4kgnyLnvZaFuGPDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T19:08:02.059017Z"},"content_sha256":"89f88ee8eafdd92c7c17daf6a463de18057ce254890699d5b461a503d1b00212","schema_version":"1.0","event_id":"sha256:89f88ee8eafdd92c7c17daf6a463de18057ce254890699d5b461a503d1b00212"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:GJ4YTDA7XX7Y6QSJ7DFIMBQS2O","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Prototypicality Bias Reveals Blindspots in Multimodal Evaluation Metrics","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Gagan Bhatia, Steffen Eger, Subhadeep Roy","submitted_at":"2026-01-08T13:49:14Z","abstract_excerpt":"Automatic metrics are widely used to evaluate text-to-image models, often replacing human judgment in benchmarking, model selection, and large-scale data filtering. Yet they may reward images that look plausible or prototypical rather than images that faithfully satisfy the prompt. We identify prototypicality bias as a systematic blindspot in multimodal evaluation: metrics can prefer a semantically incorrect but visually or socially prototypical image over a correct but less prototypical one. We introduce PROTOBIAS, a controlled diagnostic benchmark across Animals, Objects, and Demography, whe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.04946","kind":"arxiv","version":3},"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/2601.04946/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":"2026-06-02T03:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2qGk8Icm+sBuCeF2whWI95w4cz9Hiu6aETl142Qy1YhWrDfqF7dImIJ+P6PhN2AI309Dq0YtiInJM3Zw7CZXDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T19:08:02.059402Z"},"content_sha256":"57f24d47c896c6e4ddf3e7017e8d4f6e9988c13445a47160609058c0404e04f6","schema_version":"1.0","event_id":"sha256:57f24d47c896c6e4ddf3e7017e8d4f6e9988c13445a47160609058c0404e04f6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GJ4YTDA7XX7Y6QSJ7DFIMBQS2O/bundle.json","state_url":"https://pith.science/pith/GJ4YTDA7XX7Y6QSJ7DFIMBQS2O/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GJ4YTDA7XX7Y6QSJ7DFIMBQS2O/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":"2026-06-03T19:08:02Z","links":{"resolver":"https://pith.science/pith/GJ4YTDA7XX7Y6QSJ7DFIMBQS2O","bundle":"https://pith.science/pith/GJ4YTDA7XX7Y6QSJ7DFIMBQS2O/bundle.json","state":"https://pith.science/pith/GJ4YTDA7XX7Y6QSJ7DFIMBQS2O/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GJ4YTDA7XX7Y6QSJ7DFIMBQS2O/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GJ4YTDA7XX7Y6QSJ7DFIMBQS2O","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":"e1a3df6b302157012e13eb6c3a01aca6eaa5d69c5f7f2c6cdb843e3903469d32","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-01-08T13:49:14Z","title_canon_sha256":"4e5bd205d625ef46a2157ee876c341c44ba59b4adeb09d55b77a67b83191707f"},"schema_version":"1.0","source":{"id":"2601.04946","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.04946","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2601.04946v3","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.04946","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"GJ4YTDA7XX7Y","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"GJ4YTDA7XX7Y6QSJ","created_at":"2026-06-02T03:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"GJ4YTDA7","created_at":"2026-06-02T03:04:36Z"}],"graph_snapshots":[{"event_id":"sha256:57f24d47c896c6e4ddf3e7017e8d4f6e9988c13445a47160609058c0404e04f6","target":"graph","created_at":"2026-06-02T03:04:36Z","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/2601.04946/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Automatic metrics are widely used to evaluate text-to-image models, often replacing human judgment in benchmarking, model selection, and large-scale data filtering. Yet they may reward images that look plausible or prototypical rather than images that faithfully satisfy the prompt. We identify prototypicality bias as a systematic blindspot in multimodal evaluation: metrics can prefer a semantically incorrect but visually or socially prototypical image over a correct but less prototypical one. We introduce PROTOBIAS, a controlled diagnostic benchmark across Animals, Objects, and Demography, whe","authors_text":"Gagan Bhatia, Steffen Eger, Subhadeep Roy","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-01-08T13:49:14Z","title":"Prototypicality Bias Reveals Blindspots in Multimodal Evaluation Metrics"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.04946","kind":"arxiv","version":3},"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":"sha256:89f88ee8eafdd92c7c17daf6a463de18057ce254890699d5b461a503d1b00212","target":"record","created_at":"2026-06-02T03:04:36Z","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":"e1a3df6b302157012e13eb6c3a01aca6eaa5d69c5f7f2c6cdb843e3903469d32","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-01-08T13:49:14Z","title_canon_sha256":"4e5bd205d625ef46a2157ee876c341c44ba59b4adeb09d55b77a67b83191707f"},"schema_version":"1.0","source":{"id":"2601.04946","kind":"arxiv","version":3}},"canonical_sha256":"3279898c1fbdff8f4249f8ca860612d38b42e5b180faa061004acb883b134554","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3279898c1fbdff8f4249f8ca860612d38b42e5b180faa061004acb883b134554","first_computed_at":"2026-06-02T03:04:36.872044Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T03:04:36.872044Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"LmKBfz5ZS0y78OCHqMaudlmF4qkeepBmKSCekmQ63XlzK/HPTNaEYhP22UL4AyBW6nvfYgC6aNt7r/jsu+ZBAw==","signature_status":"signed_v1","signed_at":"2026-06-02T03:04:36.872555Z","signed_message":"canonical_sha256_bytes"},"source_id":"2601.04946","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:89f88ee8eafdd92c7c17daf6a463de18057ce254890699d5b461a503d1b00212","sha256:57f24d47c896c6e4ddf3e7017e8d4f6e9988c13445a47160609058c0404e04f6"],"state_sha256":"ed46285b5ff863357c7ba3d4e0124c10cb5fc685878b7601fb4ad503f202c9ee"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/s5V7jnu0jJ9AaY4zPLNHv9m2U7kWNPRRlE148r3Q+sUTlHpMyW1x+GHjy+T9WoYMIaezAmBjjeeG3DkB8dhBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-03T19:08:02.061591Z","bundle_sha256":"bb68c1b124e21540db97e5d366b3fe95961687fc747496ef01bdf9100ab6bd69"}}