{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:TF3R5ZCYI7BKZ2EWT5ZZKNGK37","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":"d28865b81ca6d8e943680b844643642480ad6d359b91c59d0bbcdd14d19e7057","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-17T22:31:10Z","title_canon_sha256":"60972c4844a245d2f15a2adb50a2178586149a2674b1eac7632f9e78ba36d925"},"schema_version":"1.0","source":{"id":"2606.19636","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.19636","created_at":"2026-06-19T16:12:30Z"},{"alias_kind":"arxiv_version","alias_value":"2606.19636v1","created_at":"2026-06-19T16:12:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.19636","created_at":"2026-06-19T16:12:30Z"},{"alias_kind":"pith_short_12","alias_value":"TF3R5ZCYI7BK","created_at":"2026-06-19T16:12:30Z"},{"alias_kind":"pith_short_16","alias_value":"TF3R5ZCYI7BKZ2EW","created_at":"2026-06-19T16:12:30Z"},{"alias_kind":"pith_short_8","alias_value":"TF3R5ZCY","created_at":"2026-06-19T16:12:30Z"}],"graph_snapshots":[{"event_id":"sha256:d52824d4b8410aef0b7fcdb88b581ccdef151655eafaa13efcc6278d26c3e022","target":"graph","created_at":"2026-06-19T16:12:30Z","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/2606.19636/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Math and science reasoning benchmarks rely on pass@k, the fraction of sampled chains that reach gold, as the canonical per-example difficulty signal. The same signal drives RL with verifiable rewards, math data curation, synthetic curricula, and verifier training. We show this proxy has a persistent blind spot on its hardest stratum: on the eight free-form math cells we test (GSM8K and MATH across four open-weight models), 10.3-22.9% of the examples that no sampling seed solves in six tries are instead solved at matched compute by a six-chain deterministic regime. These are greedy decoding plu","authors_text":"Emanuele Rodol\\`a, Luca Zhou, Roberto Dess\\`i, Sajel Shah","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-17T22:31:10Z","title":"Hard or Just Unreached? Diagnosing the Sampling Blind Spot in Math-Reasoning Difficulty Estimation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.19636","kind":"arxiv","version":1},"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":"sha256:967ec7060c7a8f5f68e34b29ea367819c12dc7186a41e142134effe6c0aca654","target":"record","created_at":"2026-06-19T16:12:30Z","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":"d28865b81ca6d8e943680b844643642480ad6d359b91c59d0bbcdd14d19e7057","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-17T22:31:10Z","title_canon_sha256":"60972c4844a245d2f15a2adb50a2178586149a2674b1eac7632f9e78ba36d925"},"schema_version":"1.0","source":{"id":"2606.19636","kind":"arxiv","version":1}},"canonical_sha256":"99771ee45847c2ace8969f739534cadff8a3aed8956119e3184ca1cac0398e99","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"99771ee45847c2ace8969f739534cadff8a3aed8956119e3184ca1cac0398e99","first_computed_at":"2026-06-19T16:12:30.958425Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-19T16:12:30.958425Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SXlguZLNb+ZEPVw9vImFYhjNt+JiUwPfYLFN2E4vWGhgWFVhetheBNJNUCDiQmoGmWvYBi2zJqzikp6aHKBeCg==","signature_status":"signed_v1","signed_at":"2026-06-19T16:12:30.958769Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.19636","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:967ec7060c7a8f5f68e34b29ea367819c12dc7186a41e142134effe6c0aca654","sha256:d52824d4b8410aef0b7fcdb88b581ccdef151655eafaa13efcc6278d26c3e022"],"state_sha256":"e6014a0cada2b53f45931c826edabc101a4a891a11aa58ab02281f74191295b9"}