{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:UTDZYGJYQXLO4RJA7FECMEV7BB","short_pith_number":"pith:UTDZYGJY","canonical_record":{"source":{"id":"2410.15490","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-20T20:07:36Z","cross_cats_sorted":["cs.MA"],"title_canon_sha256":"dfbbf81b9de7fc162817c25abeac4b970ae3e6e6a07ceab0b9f46b2a5b6a4487","abstract_canon_sha256":"2574322d1a2cd6f1dfaf4f74c4e8cce8063ad4c4f81478fe44e117f4845a9fbe"},"schema_version":"1.0"},"canonical_sha256":"a4c79c193885d6ee4520f9482612bf087d8e14fc244c034014137e6800040513","source":{"kind":"arxiv","id":"2410.15490","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.15490","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"arxiv_version","alias_value":"2410.15490v3","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.15490","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"pith_short_12","alias_value":"UTDZYGJYQXLO","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"pith_short_16","alias_value":"UTDZYGJYQXLO4RJA","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"pith_short_8","alias_value":"UTDZYGJY","created_at":"2026-07-05T09:39:28Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:UTDZYGJYQXLO4RJA7FECMEV7BB","target":"record","payload":{"canonical_record":{"source":{"id":"2410.15490","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-20T20:07:36Z","cross_cats_sorted":["cs.MA"],"title_canon_sha256":"dfbbf81b9de7fc162817c25abeac4b970ae3e6e6a07ceab0b9f46b2a5b6a4487","abstract_canon_sha256":"2574322d1a2cd6f1dfaf4f74c4e8cce8063ad4c4f81478fe44e117f4845a9fbe"},"schema_version":"1.0"},"canonical_sha256":"a4c79c193885d6ee4520f9482612bf087d8e14fc244c034014137e6800040513","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:39:28.170058Z","signature_b64":"HnF/MhJFKRqG6UFhLROOuVu8qafjS4IlzY/YMl6nT4ydyxA9tPG1W2DLodwElizzgp4QxWVNMLqQowlLtn5ECg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a4c79c193885d6ee4520f9482612bf087d8e14fc244c034014137e6800040513","last_reissued_at":"2026-07-05T09:39:28.169579Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:39:28.169579Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.15490","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:39:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9svUoMR/fu+hLp5K3Qsxy0j2mwM0fTwXPeznxekq1NJoyZ9eV4A+vWNoIifzVo+nZtBrheZb1s4CWjjhiJLRCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T08:01:25.505953Z"},"content_sha256":"7c77f0a16e117caae842ceec2d6631572d9b5007c65ed42dc8f58d6a72f4d537","schema_version":"1.0","event_id":"sha256:7c77f0a16e117caae842ceec2d6631572d9b5007c65ed42dc8f58d6a72f4d537"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:UTDZYGJYQXLO4RJA7FECMEV7BB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Dynamic Intelligence Assessment: Benchmarking LLMs on the Road to AGI with a Focus on Model Confidence","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.MA"],"primary_cat":"cs.AI","authors_text":"Audun Josang, Bertalan Borsos, Bilel Cherif, Lajos Muzsai, Lucas C. Cordeiro, Merouane Debbah, Mohamed Amine Ferrag, Norbert Tihanyi, Rebeka Toth, Richard A. Dubniczky, Ridhi Jain, Ryan Marinelli, Tamas Bisztray, Vasileios Mavroeidis","submitted_at":"2024-10-20T20:07:36Z","abstract_excerpt":"As machine intelligence evolves, the need to test and compare the problem-solving abilities of different AI models grows. However, current benchmarks are often simplistic, allowing models to perform uniformly well and making it difficult to distinguish their capabilities. Additionally, benchmarks typically rely on static question-answer pairs that the models might memorize or guess. To address these limitations, we introduce Dynamic Intelligence Assessment (DIA), a novel methodology for testing AI models using dynamic question templates and improved metrics across multiple disciplines such as "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.15490","kind":"arxiv","version":3},"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/2410.15490/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:39:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8t3y2r92m4weDYfQpATNZVkzeo3lyFofn5jUpUbD22x+LAaDi8l8OY32je7v1jzFrRvDFQ7+z8tW/nSShHDmBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T08:01:25.506426Z"},"content_sha256":"a54dc3847938dc716af26c3bad613f745775bee89ad8d3d8f9b520c3f18fe7d5","schema_version":"1.0","event_id":"sha256:a54dc3847938dc716af26c3bad613f745775bee89ad8d3d8f9b520c3f18fe7d5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/UTDZYGJYQXLO4RJA7FECMEV7BB/bundle.json","state_url":"https://pith.science/pith/UTDZYGJYQXLO4RJA7FECMEV7BB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/UTDZYGJYQXLO4RJA7FECMEV7BB/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-04T08:01:25Z","links":{"resolver":"https://pith.science/pith/UTDZYGJYQXLO4RJA7FECMEV7BB","bundle":"https://pith.science/pith/UTDZYGJYQXLO4RJA7FECMEV7BB/bundle.json","state":"https://pith.science/pith/UTDZYGJYQXLO4RJA7FECMEV7BB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/UTDZYGJYQXLO4RJA7FECMEV7BB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:UTDZYGJYQXLO4RJA7FECMEV7BB","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":"2574322d1a2cd6f1dfaf4f74c4e8cce8063ad4c4f81478fe44e117f4845a9fbe","cross_cats_sorted":["cs.MA"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-20T20:07:36Z","title_canon_sha256":"dfbbf81b9de7fc162817c25abeac4b970ae3e6e6a07ceab0b9f46b2a5b6a4487"},"schema_version":"1.0","source":{"id":"2410.15490","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.15490","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"arxiv_version","alias_value":"2410.15490v3","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.15490","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"pith_short_12","alias_value":"UTDZYGJYQXLO","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"pith_short_16","alias_value":"UTDZYGJYQXLO4RJA","created_at":"2026-07-05T09:39:28Z"},{"alias_kind":"pith_short_8","alias_value":"UTDZYGJY","created_at":"2026-07-05T09:39:28Z"}],"graph_snapshots":[{"event_id":"sha256:a54dc3847938dc716af26c3bad613f745775bee89ad8d3d8f9b520c3f18fe7d5","target":"graph","created_at":"2026-07-05T09:39:28Z","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/2410.15490/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As machine intelligence evolves, the need to test and compare the problem-solving abilities of different AI models grows. However, current benchmarks are often simplistic, allowing models to perform uniformly well and making it difficult to distinguish their capabilities. Additionally, benchmarks typically rely on static question-answer pairs that the models might memorize or guess. To address these limitations, we introduce Dynamic Intelligence Assessment (DIA), a novel methodology for testing AI models using dynamic question templates and improved metrics across multiple disciplines such as ","authors_text":"Audun Josang, Bertalan Borsos, Bilel Cherif, Lajos Muzsai, Lucas C. Cordeiro, Merouane Debbah, Mohamed Amine Ferrag, Norbert Tihanyi, Rebeka Toth, Richard A. Dubniczky, Ridhi Jain, Ryan Marinelli, Tamas Bisztray, Vasileios Mavroeidis","cross_cats":["cs.MA"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-20T20:07:36Z","title":"Dynamic Intelligence Assessment: Benchmarking LLMs on the Road to AGI with a Focus on Model Confidence"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.15490","kind":"arxiv","version":3},"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":"sha256:7c77f0a16e117caae842ceec2d6631572d9b5007c65ed42dc8f58d6a72f4d537","target":"record","created_at":"2026-07-05T09:39:28Z","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":"2574322d1a2cd6f1dfaf4f74c4e8cce8063ad4c4f81478fe44e117f4845a9fbe","cross_cats_sorted":["cs.MA"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-20T20:07:36Z","title_canon_sha256":"dfbbf81b9de7fc162817c25abeac4b970ae3e6e6a07ceab0b9f46b2a5b6a4487"},"schema_version":"1.0","source":{"id":"2410.15490","kind":"arxiv","version":3}},"canonical_sha256":"a4c79c193885d6ee4520f9482612bf087d8e14fc244c034014137e6800040513","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a4c79c193885d6ee4520f9482612bf087d8e14fc244c034014137e6800040513","first_computed_at":"2026-07-05T09:39:28.169579Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:39:28.169579Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HnF/MhJFKRqG6UFhLROOuVu8qafjS4IlzY/YMl6nT4ydyxA9tPG1W2DLodwElizzgp4QxWVNMLqQowlLtn5ECg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:39:28.170058Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.15490","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7c77f0a16e117caae842ceec2d6631572d9b5007c65ed42dc8f58d6a72f4d537","sha256:a54dc3847938dc716af26c3bad613f745775bee89ad8d3d8f9b520c3f18fe7d5"],"state_sha256":"317633067e8dbe2b4d4c733a1b6e79d733e6fc3fc48afa48606050c5bd9b5896"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9TwVc4vxxfiW8GBGxOf0YLLABc8WuQhZCOnF6E7Z9exXoFLpXxTi9dy3iEIAlJJ5AVzVrDDaGyLUk2ZVWx1NBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T08:01:25.510962Z","bundle_sha256":"dc70d81b6520ce282928b03748bd01f104d1a05394e4acdac40e18bbde587293"}}