{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4B5WMF725OVD6FUYFB7P2QK6WB","short_pith_number":"pith:4B5WMF72","schema_version":"1.0","canonical_sha256":"e07b6617faebaa3f1698287efd415eb07152a229e02a2863720efb48cf61a4c7","source":{"kind":"arxiv","id":"2507.00432","version":2},"attestation_state":"computed","paper":{"title":"Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Graham Neubig, Maggie Huan, Minxin Du, Radha Poovendran, Seungone Kim, Tuney Zheng, Xiang Yue, Xiaoyu Xu, Yuetai Li","submitted_at":"2025-07-01T05:23:05Z","abstract_excerpt":"Math reasoning has become the poster child of progress in large language models (LLMs), with new models rapidly surpassing human-level performance on benchmarks like MATH and AIME. But as math leaderboards improve week by week, it is worth asking: do these gains reflect broader problem-solving ability or just narrow overfitting? To answer this question, we evaluate over 20 open-weight reasoning-tuned models across a broad suite of tasks, including math, scientific QA, agent planning, coding, and standard instruction-following. We surprisingly find that most models that succeed in math fail to "},"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":"2507.00432","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-07-01T05:23:05Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"336fa8cf47311762979f709847d68feaae704b54ee009a90a7d886f730c9f106","abstract_canon_sha256":"00e49dba5ded089212ca29a8316090a26c77a71120d21d8e1fb4f3782c9b1d57"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-19T00:54:51.301064Z","signature_b64":"T042PHwYLa+bLXoIz24XWUob1sj3M3g+KCgS7afDizyr2opts8MK5zExWKe924gHAFo8+uPEEZXtTzkqrygoCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e07b6617faebaa3f1698287efd415eb07152a229e02a2863720efb48cf61a4c7","last_reissued_at":"2026-05-19T00:54:51.297971Z","signature_status":"signed_v1","first_computed_at":"2026-05-19T00:54:51.297971Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Graham Neubig, Maggie Huan, Minxin Du, Radha Poovendran, Seungone Kim, Tuney Zheng, Xiang Yue, Xiaoyu Xu, Yuetai Li","submitted_at":"2025-07-01T05:23:05Z","abstract_excerpt":"Math reasoning has become the poster child of progress in large language models (LLMs), with new models rapidly surpassing human-level performance on benchmarks like MATH and AIME. But as math leaderboards improve week by week, it is worth asking: do these gains reflect broader problem-solving ability or just narrow overfitting? To answer this question, we evaluate over 20 open-weight reasoning-tuned models across a broad suite of tasks, including math, scientific QA, agent planning, coding, and standard instruction-following. We surprisingly find that most models that succeed in math fail to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.00432","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":""},"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":"2507.00432","created_at":"2026-05-19T00:54:51.298152+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.00432v2","created_at":"2026-05-19T00:54:51.298152+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.00432","created_at":"2026-05-19T00:54:51.298152+00:00"},{"alias_kind":"pith_short_12","alias_value":"4B5WMF725OVD","created_at":"2026-05-19T00:54:51.298152+00:00"},{"alias_kind":"pith_short_16","alias_value":"4B5WMF725OVD6FUY","created_at":"2026-05-19T00:54:51.298152+00:00"},{"alias_kind":"pith_short_8","alias_value":"4B5WMF72","created_at":"2026-05-19T00:54:51.298152+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":18,"sample":[{"citing_arxiv_id":"2508.06412","citing_title":"Sample-efficient LLM Optimization with Reset Replay","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2508.17784","citing_title":"Proximal Supervised Fine-Tuning","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2508.20697","citing_title":"Token Buncher: Shielding LLMs from Harmful Reinforcement Learning Fine-Tuning","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2509.08827","citing_title":"A Survey of Reinforcement Learning for Large Reasoning Models","ref_index":208,"is_internal_anchor":true},{"citing_arxiv_id":"2511.18787","citing_title":"Understanding Task Transfer in Vision-Language Models","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2512.11470","citing_title":"Rethinking Expert Trajectory Utilization in LLM Post-training for Mathematical Reasoning","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2601.07389","citing_title":"On the Non-decoupling of Supervised Fine-tuning and Reinforcement Learning in Post-training","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2601.17467","citing_title":"Harnessing Reasoning Trajectories for Hallucination Detection via Answer-agreement Representation Shaping","ref_index":15,"is_internal_anchor":true},{"citing_arxiv_id":"2601.21484","citing_title":"ETS: Energy-Guided Test-Time Scaling for Training-Free RL Alignment","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10973","citing_title":"Rotation-Preserving Supervised Fine-Tuning","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11388","citing_title":"Deep Reasoning in General Purpose Agents via Structured Meta-Cognition","ref_index":153,"is_internal_anchor":true},{"citing_arxiv_id":"2503.09567","citing_title":"Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models","ref_index":281,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09879","citing_title":"M2A: Synergizing Mathematical and Agentic Reasoning in Large Language Models","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2601.18734","citing_title":"Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2604.25011","citing_title":"Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2604.08477","citing_title":"SUPERNOVA: Eliciting General Reasoning in LLMs with Reinforcement Learning on Natural Instructions","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2604.17614","citing_title":"Characterizing Model-Native Skills","ref_index":79,"is_internal_anchor":true},{"citing_arxiv_id":"2604.17928","citing_title":"HEALing Entropy Collapse: Enhancing Exploration in Few-Shot RLVR via Hybrid-Domain Entropy Dynamics Alignment","ref_index":46,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB","json":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB.json","graph_json":"https://pith.science/api/pith-number/4B5WMF725OVD6FUYFB7P2QK6WB/graph.json","events_json":"https://pith.science/api/pith-number/4B5WMF725OVD6FUYFB7P2QK6WB/events.json","paper":"https://pith.science/paper/4B5WMF72"},"agent_actions":{"view_html":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB","download_json":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB.json","view_paper":"https://pith.science/paper/4B5WMF72","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.00432&json=true","fetch_graph":"https://pith.science/api/pith-number/4B5WMF725OVD6FUYFB7P2QK6WB/graph.json","fetch_events":"https://pith.science/api/pith-number/4B5WMF725OVD6FUYFB7P2QK6WB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB/action/storage_attestation","attest_author":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB/action/author_attestation","sign_citation":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB/action/citation_signature","submit_replication":"https://pith.science/pith/4B5WMF725OVD6FUYFB7P2QK6WB/action/replication_record"}},"created_at":"2026-05-19T00:54:51.298152+00:00","updated_at":"2026-05-19T00:54:51.298152+00:00"}