{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:D4MHOGRRERP53YAI5JXVRM3IWB","short_pith_number":"pith:D4MHOGRR","schema_version":"1.0","canonical_sha256":"1f18771a31245fdde008ea6f58b368b069d201ccbfc54fe6b0feaae596d34428","source":{"kind":"arxiv","id":"2401.05604","version":2},"attestation_state":"computed","paper":{"title":"REBUS: A Robust Evaluation Benchmark of Understanding Symbols","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.CY"],"primary_cat":"cs.CL","authors_text":"Aaron Kirtland, Andrew Gritsevskiy, Arjun Panickssery, Derik Kauffman, Hans Gundlach, Irina Gritsevskaya, Joe Cavanagh, Jonathan Chiang, Lydia La Roux, Michelle Hung","submitted_at":"2024-01-11T00:30:28Z","abstract_excerpt":"We propose a new benchmark evaluating the performance of multimodal large language models on rebus puzzles. The dataset covers 333 original examples of image-based wordplay, cluing 13 categories such as movies, composers, major cities, and food. To achieve good performance on the benchmark of identifying the clued word or phrase, models must combine image recognition and string manipulation with hypothesis testing, multi-step reasoning, and an understanding of human cognition, making for a complex, multimodal evaluation of capabilities. We find that GPT-4o significantly outperforms all other m"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2401.05604","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-11T00:30:28Z","cross_cats_sorted":["cs.AI","cs.CV","cs.CY"],"title_canon_sha256":"e0ffce1776d4f7762a3f2e6b8dbe26288865c506a8ed320ece670708afe86c4f","abstract_canon_sha256":"2a2c3c1ebd03e79da806bfbaa734cf138cfd6c7f1fd64448eaf7e241ff20ac04"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:26:43.385801Z","signature_b64":"B4Yud5hnXm1xjh2YM/d5raG2VIJXtoH4uEVy9TH6Ftiff6ZQRmTkWdX5PLANOHJjkFp+Q3hKSblf9gzpwNSjAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1f18771a31245fdde008ea6f58b368b069d201ccbfc54fe6b0feaae596d34428","last_reissued_at":"2026-07-05T08:26:43.385270Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:26:43.385270Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"REBUS: A Robust Evaluation Benchmark of Understanding Symbols","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.CY"],"primary_cat":"cs.CL","authors_text":"Aaron Kirtland, Andrew Gritsevskiy, Arjun Panickssery, Derik Kauffman, Hans Gundlach, Irina Gritsevskaya, Joe Cavanagh, Jonathan Chiang, Lydia La Roux, Michelle Hung","submitted_at":"2024-01-11T00:30:28Z","abstract_excerpt":"We propose a new benchmark evaluating the performance of multimodal large language models on rebus puzzles. The dataset covers 333 original examples of image-based wordplay, cluing 13 categories such as movies, composers, major cities, and food. To achieve good performance on the benchmark of identifying the clued word or phrase, models must combine image recognition and string manipulation with hypothesis testing, multi-step reasoning, and an understanding of human cognition, making for a complex, multimodal evaluation of capabilities. We find that GPT-4o significantly outperforms all other m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.05604","kind":"arxiv","version":2},"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/2401.05604/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2401.05604","created_at":"2026-07-05T08:26:43.385341+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.05604v2","created_at":"2026-07-05T08:26:43.385341+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.05604","created_at":"2026-07-05T08:26:43.385341+00:00"},{"alias_kind":"pith_short_12","alias_value":"D4MHOGRRERP5","created_at":"2026-07-05T08:26:43.385341+00:00"},{"alias_kind":"pith_short_16","alias_value":"D4MHOGRRERP53YAI","created_at":"2026-07-05T08:26:43.385341+00:00"},{"alias_kind":"pith_short_8","alias_value":"D4MHOGRR","created_at":"2026-07-05T08:26:43.385341+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.01619","citing_title":"Throttling Web Agents Using Reasoning Gates","ref_index":37,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB","json":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB.json","graph_json":"https://pith.science/api/pith-number/D4MHOGRRERP53YAI5JXVRM3IWB/graph.json","events_json":"https://pith.science/api/pith-number/D4MHOGRRERP53YAI5JXVRM3IWB/events.json","paper":"https://pith.science/paper/D4MHOGRR"},"agent_actions":{"view_html":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB","download_json":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB.json","view_paper":"https://pith.science/paper/D4MHOGRR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.05604&json=true","fetch_graph":"https://pith.science/api/pith-number/D4MHOGRRERP53YAI5JXVRM3IWB/graph.json","fetch_events":"https://pith.science/api/pith-number/D4MHOGRRERP53YAI5JXVRM3IWB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB/action/storage_attestation","attest_author":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB/action/author_attestation","sign_citation":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB/action/citation_signature","submit_replication":"https://pith.science/pith/D4MHOGRRERP53YAI5JXVRM3IWB/action/replication_record"}},"created_at":"2026-07-05T08:26:43.385341+00:00","updated_at":"2026-07-05T08:26:43.385341+00:00"}