{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:AOZJ4KMLULKVUY6WUFPB7CJ4A2","short_pith_number":"pith:AOZJ4KML","schema_version":"1.0","canonical_sha256":"03b29e298ba2d55a63d6a15e1f893c0683b6cc32cbe4acfa784e6a6160d4ae61","source":{"kind":"arxiv","id":"2407.21009","version":4},"attestation_state":"computed","paper":{"title":"AI-Assisted Generation of Difficult Math Questions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Anirudh Goyal, Dingli Yu, Jiatong Yu, Kaifeng Lyu, Michael Mozer, Nan Rosemary Ke, Sanjeev Arora, Simon Park, Vedant Shah, Yinghui He, Yoshua Bengio","submitted_at":"2024-07-30T17:55:36Z","abstract_excerpt":"Current LLM training positions mathematical reasoning as a core capability. With publicly available sources fully tapped, there is unmet demand for diverse and challenging math questions. Relying solely on human experts is both time-consuming and costly, while LLM-generated questions often lack the requisite diversity and difficulty. We present a design framework that combines the strengths of LLMs with a human-in-the-loop approach to generate a diverse array of challenging math questions. We leverage LLM metacognition skills [Didolkar et al., 2024] of a strong LLM to extract core \"skills\" fro"},"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":"2407.21009","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-07-30T17:55:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"343c909bc6387a3e7b7ade2224bf7cc7eb8e04d9a9222361ad233e92349c5d42","abstract_canon_sha256":"533ab721c982e7dcd399537c8fb5b9fa436c4a3f450b1b4cbe257e17503e391e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:08:19.786622Z","signature_b64":"MGRowunD4KQiaCBArV/003b2kiyE4c4DsuZUrIE7ZGlpRu1m3jCdHPrQ53YCdKVZco+75XF/Vr+VJ6uPvXvdBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03b29e298ba2d55a63d6a15e1f893c0683b6cc32cbe4acfa784e6a6160d4ae61","last_reissued_at":"2026-07-05T10:08:19.786053Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:08:19.786053Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AI-Assisted Generation of Difficult Math Questions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Anirudh Goyal, Dingli Yu, Jiatong Yu, Kaifeng Lyu, Michael Mozer, Nan Rosemary Ke, Sanjeev Arora, Simon Park, Vedant Shah, Yinghui He, Yoshua Bengio","submitted_at":"2024-07-30T17:55:36Z","abstract_excerpt":"Current LLM training positions mathematical reasoning as a core capability. With publicly available sources fully tapped, there is unmet demand for diverse and challenging math questions. Relying solely on human experts is both time-consuming and costly, while LLM-generated questions often lack the requisite diversity and difficulty. We present a design framework that combines the strengths of LLMs with a human-in-the-loop approach to generate a diverse array of challenging math questions. We leverage LLM metacognition skills [Didolkar et al., 2024] of a strong LLM to extract core \"skills\" fro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.21009","kind":"arxiv","version":4},"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/2407.21009/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":"2407.21009","created_at":"2026-07-05T10:08:19.786130+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.21009v4","created_at":"2026-07-05T10:08:19.786130+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.21009","created_at":"2026-07-05T10:08:19.786130+00:00"},{"alias_kind":"pith_short_12","alias_value":"AOZJ4KMLULKV","created_at":"2026-07-05T10:08:19.786130+00:00"},{"alias_kind":"pith_short_16","alias_value":"AOZJ4KMLULKVUY6W","created_at":"2026-07-05T10:08:19.786130+00:00"},{"alias_kind":"pith_short_8","alias_value":"AOZJ4KML","created_at":"2026-07-05T10:08:19.786130+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25996","citing_title":"Autodata: An agentic data scientist to create high quality synthetic data","ref_index":109,"is_internal_anchor":false},{"citing_arxiv_id":"2606.25996","citing_title":"Autodata: An agentic data scientist to create high quality synthetic data","ref_index":109,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21972","citing_title":"Human vs Machine Mathematical Difficulty on Project Euler: An Experimental Analysis","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04386","citing_title":"Automatically Generating Hard Math Problems from Hypothesis-Driven Error Analysis","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2","json":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2.json","graph_json":"https://pith.science/api/pith-number/AOZJ4KMLULKVUY6WUFPB7CJ4A2/graph.json","events_json":"https://pith.science/api/pith-number/AOZJ4KMLULKVUY6WUFPB7CJ4A2/events.json","paper":"https://pith.science/paper/AOZJ4KML"},"agent_actions":{"view_html":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2","download_json":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2.json","view_paper":"https://pith.science/paper/AOZJ4KML","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.21009&json=true","fetch_graph":"https://pith.science/api/pith-number/AOZJ4KMLULKVUY6WUFPB7CJ4A2/graph.json","fetch_events":"https://pith.science/api/pith-number/AOZJ4KMLULKVUY6WUFPB7CJ4A2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2/action/storage_attestation","attest_author":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2/action/author_attestation","sign_citation":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2/action/citation_signature","submit_replication":"https://pith.science/pith/AOZJ4KMLULKVUY6WUFPB7CJ4A2/action/replication_record"}},"created_at":"2026-07-05T10:08:19.786130+00:00","updated_at":"2026-07-05T10:08:19.786130+00:00"}