{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:AR6UHUZYQGJVTCEDYJKF4E4GZE","short_pith_number":"pith:AR6UHUZY","schema_version":"1.0","canonical_sha256":"047d43d3388193598883c2545e1386c902d69fc986c273254a45320e33a70e66","source":{"kind":"arxiv","id":"2606.07515","version":1},"attestation_state":"computed","paper":{"title":"How reliable are LLMs when it comes to playing dice?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.HC","math.PR"],"primary_cat":"cs.CL","authors_text":"Bernardo Busoni, Gianmarco Bet, Luca Avena","submitted_at":"2026-06-05T17:59:42Z","abstract_excerpt":"We investigate the probabilistic reasoning capabilities of large language models through a controlled benchmarking study on discrete probability problems. We constructed two datasets, respectively a set of standard exercises and a set of counterintuitive exercises, designed to trigger heuristic reasoning, and evaluated 8 state-of-the-art models, each tested with and without Chain-of-Thought prompting. Models achieve an average accuracy of 0.96 on standard problems but only 0.59 on counterintuitive ones. We further provide empirical evidence of token bias: performance drops by over 20% when can"},"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":"2606.07515","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T17:59:42Z","cross_cats_sorted":["cs.AI","cs.HC","math.PR"],"title_canon_sha256":"37863ae2f143b3e4c150bd2ec951a041a91545e6dab0bf83f04fc1dc663f3710","abstract_canon_sha256":"01e0c24cfc6c9e3e52479773a211f8a65c85cb492c7cce71868e3ce1c8e21fc8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:05:32.297477Z","signature_b64":"i9CCKAPfqnk/JiHsUPHgqj+EvTjmM2GwNROKKymMVzYDFglwdqgtrxHQltWiKEKcq6Ea1UBN1ZSdqA9ZOAGPAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"047d43d3388193598883c2545e1386c902d69fc986c273254a45320e33a70e66","last_reissued_at":"2026-06-08T01:05:32.296552Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:05:32.296552Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How reliable are LLMs when it comes to playing dice?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.HC","math.PR"],"primary_cat":"cs.CL","authors_text":"Bernardo Busoni, Gianmarco Bet, Luca Avena","submitted_at":"2026-06-05T17:59:42Z","abstract_excerpt":"We investigate the probabilistic reasoning capabilities of large language models through a controlled benchmarking study on discrete probability problems. We constructed two datasets, respectively a set of standard exercises and a set of counterintuitive exercises, designed to trigger heuristic reasoning, and evaluated 8 state-of-the-art models, each tested with and without Chain-of-Thought prompting. Models achieve an average accuracy of 0.96 on standard problems but only 0.59 on counterintuitive ones. We further provide empirical evidence of token bias: performance drops by over 20% when can"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07515","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/2606.07515/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":"2606.07515","created_at":"2026-06-08T01:05:32.296723+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.07515v1","created_at":"2026-06-08T01:05:32.296723+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07515","created_at":"2026-06-08T01:05:32.296723+00:00"},{"alias_kind":"pith_short_12","alias_value":"AR6UHUZYQGJV","created_at":"2026-06-08T01:05:32.296723+00:00"},{"alias_kind":"pith_short_16","alias_value":"AR6UHUZYQGJVTCED","created_at":"2026-06-08T01:05:32.296723+00:00"},{"alias_kind":"pith_short_8","alias_value":"AR6UHUZY","created_at":"2026-06-08T01:05:32.296723+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE","json":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE.json","graph_json":"https://pith.science/api/pith-number/AR6UHUZYQGJVTCEDYJKF4E4GZE/graph.json","events_json":"https://pith.science/api/pith-number/AR6UHUZYQGJVTCEDYJKF4E4GZE/events.json","paper":"https://pith.science/paper/AR6UHUZY"},"agent_actions":{"view_html":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE","download_json":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE.json","view_paper":"https://pith.science/paper/AR6UHUZY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.07515&json=true","fetch_graph":"https://pith.science/api/pith-number/AR6UHUZYQGJVTCEDYJKF4E4GZE/graph.json","fetch_events":"https://pith.science/api/pith-number/AR6UHUZYQGJVTCEDYJKF4E4GZE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE/action/storage_attestation","attest_author":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE/action/author_attestation","sign_citation":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE/action/citation_signature","submit_replication":"https://pith.science/pith/AR6UHUZYQGJVTCEDYJKF4E4GZE/action/replication_record"}},"created_at":"2026-06-08T01:05:32.296723+00:00","updated_at":"2026-06-08T01:05:32.296723+00:00"}