{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4F3AOGJSCSNYUTAN7ULQR6UPW4","short_pith_number":"pith:4F3AOGJS","schema_version":"1.0","canonical_sha256":"e176071932149b8a4c0dfd1708fa8fb73c84e75622ff87cf6d150ca2bc75f2dc","source":{"kind":"arxiv","id":"2506.18880","version":1},"attestation_state":"computed","paper":{"title":"OMEGA: Can LLMs Reason Outside the Box in Math? Evaluating Exploratory, Compositional, and Transformative Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dawn Song, Georgia Zhou, Hannaneh Hajishirzi, Ken Zheng, Nouha Dziri, Shawn Hu, Yiyou Sun","submitted_at":"2025-06-23T17:51:40Z","abstract_excerpt":"Recent large-scale language models (LLMs) with long Chain-of-Thought reasoning-such as DeepSeek-R1-have achieved impressive results on Olympiad-level mathematics benchmarks. However, they often rely on a narrow set of strategies and struggle with problems that require a novel way of thinking. To systematically investigate these limitations, we introduce OMEGA-Out-of-distribution Math Problems Evaluation with 3 Generalization Axes-a controlled yet diverse benchmark designed to evaluate three axes of out-of-distribution generalization, inspired by Boden's typology of creativity: (1) Exploratory-"},"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":"2506.18880","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-23T17:51:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e292e6a95e2258b48516750ca4002059b4f21a2bed271fbc99ab96caf580999b","abstract_canon_sha256":"ec122cf06bd4861048d686ae2d1795416a133996115bdcf39cce6ffb6532c8ab"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:57.468996Z","signature_b64":"SrDLvHTzK1GCd5OnueOjrDw6+FKgZUu/aB1riB5HN+WJusO2H5M9eEzjzT9eJ/WlLIczrIUx47AVdrtViejYAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e176071932149b8a4c0dfd1708fa8fb73c84e75622ff87cf6d150ca2bc75f2dc","last_reissued_at":"2026-07-05T11:25:57.468500Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:57.468500Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OMEGA: Can LLMs Reason Outside the Box in Math? Evaluating Exploratory, Compositional, and Transformative Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dawn Song, Georgia Zhou, Hannaneh Hajishirzi, Ken Zheng, Nouha Dziri, Shawn Hu, Yiyou Sun","submitted_at":"2025-06-23T17:51:40Z","abstract_excerpt":"Recent large-scale language models (LLMs) with long Chain-of-Thought reasoning-such as DeepSeek-R1-have achieved impressive results on Olympiad-level mathematics benchmarks. However, they often rely on a narrow set of strategies and struggle with problems that require a novel way of thinking. To systematically investigate these limitations, we introduce OMEGA-Out-of-distribution Math Problems Evaluation with 3 Generalization Axes-a controlled yet diverse benchmark designed to evaluate three axes of out-of-distribution generalization, inspired by Boden's typology of creativity: (1) Exploratory-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.18880","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/2506.18880/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":"2506.18880","created_at":"2026-07-05T11:25:57.468564+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.18880v1","created_at":"2026-07-05T11:25:57.468564+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.18880","created_at":"2026-07-05T11:25:57.468564+00:00"},{"alias_kind":"pith_short_12","alias_value":"4F3AOGJSCSNY","created_at":"2026-07-05T11:25:57.468564+00:00"},{"alias_kind":"pith_short_16","alias_value":"4F3AOGJSCSNYUTAN","created_at":"2026-07-05T11:25:57.468564+00:00"},{"alias_kind":"pith_short_8","alias_value":"4F3AOGJS","created_at":"2026-07-05T11:25:57.468564+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05992","citing_title":"PluraMath: Extending Mathematical Reasoning Evaluation Beyond High-Resource Languages","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2606.26091","citing_title":"On-Policy Self-Distillation with Sampled Demonstrations Reduces Output Diversity","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18089","citing_title":"From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning","ref_index":94,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13450","citing_title":"Assessing the Creativity of Large Language Models: Testing, Limits, and New Frontiers","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18473","citing_title":"Train Separately, Merge Together: Modular Post-Training with Mixture-of-Experts","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4","json":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4.json","graph_json":"https://pith.science/api/pith-number/4F3AOGJSCSNYUTAN7ULQR6UPW4/graph.json","events_json":"https://pith.science/api/pith-number/4F3AOGJSCSNYUTAN7ULQR6UPW4/events.json","paper":"https://pith.science/paper/4F3AOGJS"},"agent_actions":{"view_html":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4","download_json":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4.json","view_paper":"https://pith.science/paper/4F3AOGJS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.18880&json=true","fetch_graph":"https://pith.science/api/pith-number/4F3AOGJSCSNYUTAN7ULQR6UPW4/graph.json","fetch_events":"https://pith.science/api/pith-number/4F3AOGJSCSNYUTAN7ULQR6UPW4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4/action/storage_attestation","attest_author":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4/action/author_attestation","sign_citation":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4/action/citation_signature","submit_replication":"https://pith.science/pith/4F3AOGJSCSNYUTAN7ULQR6UPW4/action/replication_record"}},"created_at":"2026-07-05T11:25:57.468564+00:00","updated_at":"2026-07-05T11:25:57.468564+00:00"}