{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:73XVJFIDPLE4YCRTXMX5OJBZ3Y","short_pith_number":"pith:73XVJFID","canonical_record":{"source":{"id":"2503.06378","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-03-09T01:13:56Z","cross_cats_sorted":["cs.CL","cs.CY"],"title_canon_sha256":"78e7f8bc44dd37e6b303c5569cc1e76b709c498f708e4f6b3cb80516079f55d0","abstract_canon_sha256":"10b86c7e748898bb949194fde48700de87f56553e85d6f0ada7af2ec2df82e41"},"schema_version":"1.0"},"canonical_sha256":"feef5495037ac9cc0a33bb2fd72439de02db98d5617f48129402aa2ce87e7d48","source":{"kind":"arxiv","id":"2503.06378","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06378","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06378v2","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06378","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"pith_short_12","alias_value":"73XVJFIDPLE4","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"pith_short_16","alias_value":"73XVJFIDPLE4YCRT","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"pith_short_8","alias_value":"73XVJFID","created_at":"2026-07-05T10:31:56Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:73XVJFIDPLE4YCRTXMX5OJBZ3Y","target":"record","payload":{"canonical_record":{"source":{"id":"2503.06378","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-03-09T01:13:56Z","cross_cats_sorted":["cs.CL","cs.CY"],"title_canon_sha256":"78e7f8bc44dd37e6b303c5569cc1e76b709c498f708e4f6b3cb80516079f55d0","abstract_canon_sha256":"10b86c7e748898bb949194fde48700de87f56553e85d6f0ada7af2ec2df82e41"},"schema_version":"1.0"},"canonical_sha256":"feef5495037ac9cc0a33bb2fd72439de02db98d5617f48129402aa2ce87e7d48","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:31:56.243744Z","signature_b64":"JNvD1TvafoXQx90dfK/GLkgpDbNcb5ydxbvkIZQr96jdvTVdXj8B28aBiYewmkbMtfuntIovVIcxT3dVgJwuAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"feef5495037ac9cc0a33bb2fd72439de02db98d5617f48129402aa2ce87e7d48","last_reissued_at":"2026-07-05T10:31:56.243015Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:31:56.243015Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.06378","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:31:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5YRKrD3DgGVKcmxIn23aYi4XyRUZ0m+yKGlD3DvdzZg9RSYfZTHijX0daIWKcS+AjpOSiP60zMzcSF0xsRcuCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T00:39:25.155054Z"},"content_sha256":"69ebfe62522c56f198b5c80f5a5e58a880624732f45a1762b24e7bc486d5aff3","schema_version":"1.0","event_id":"sha256:69ebfe62522c56f198b5c80f5a5e58a880624732f45a1762b24e7bc486d5aff3"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:73XVJFIDPLE4YCRTXMX5OJBZ3Y","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"General Scales Unlock AI Evaluation with Explanatory and Predictive Power","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CY"],"primary_cat":"cs.AI","authors_text":"Behzad Mehrbakhsh, David Stillwell, Fernando Mart\\'inez-Plumed, Jindong Wang, Jiyun Zu, John Burden, Jonathan E. Prunty, Jos\\'e Hern\\'andez-Orallo, Katherine M. Collins, Kexin Jiang Chen, Lexin Zhou, Lorenzo Pacchiardi, Lucy Cheke, Luning Sun, Manuel Cebrian, Pablo A. M. Casares, Pablo S\\'anchez-Garc\\'ia, Patrick C. Kyllonen, Peter Henderson, Qinlin Zhao, Seraphina Zhang, Sherry Tongshuang Wu, Xing Xie, Yael Moros-Daval, Yitian Huang, Zongqian Li","submitted_at":"2025-03-09T01:13:56Z","abstract_excerpt":"Ensuring safe and effective use of AI requires understanding and anticipating its performance on novel tasks, from advanced scientific challenges to transformed workplace activities. So far, benchmarking has guided progress in AI, but it has offered limited explanatory and predictive power for general-purpose AI systems, given the low transferability across diverse tasks. In this paper, we introduce general scales for AI evaluation that can explain what common AI benchmarks really measure, extract ability profiles of AI systems, and predict their performance for new task instances, in- and out"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06378","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2503.06378/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":"2026-07-05T10:31:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"sBX8hOTdVjt5DkHORhQkm73JyOEeyg29P3BvNo02PB+6OI7XMW4D6VMPoVLNs/vyrkKfmuof/CNhrcuRqpocBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T00:39:25.155434Z"},"content_sha256":"3ed88b9699bb700928f1ed7a560e5ca301d83723c0f6fbcda2f2e36176780c4a","schema_version":"1.0","event_id":"sha256:3ed88b9699bb700928f1ed7a560e5ca301d83723c0f6fbcda2f2e36176780c4a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/73XVJFIDPLE4YCRTXMX5OJBZ3Y/bundle.json","state_url":"https://pith.science/pith/73XVJFIDPLE4YCRTXMX5OJBZ3Y/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/73XVJFIDPLE4YCRTXMX5OJBZ3Y/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":"2026-08-04T00:39:25Z","links":{"resolver":"https://pith.science/pith/73XVJFIDPLE4YCRTXMX5OJBZ3Y","bundle":"https://pith.science/pith/73XVJFIDPLE4YCRTXMX5OJBZ3Y/bundle.json","state":"https://pith.science/pith/73XVJFIDPLE4YCRTXMX5OJBZ3Y/state.json","well_known_bundle":"https://pith.science/.well-known/pith/73XVJFIDPLE4YCRTXMX5OJBZ3Y/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:73XVJFIDPLE4YCRTXMX5OJBZ3Y","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":"10b86c7e748898bb949194fde48700de87f56553e85d6f0ada7af2ec2df82e41","cross_cats_sorted":["cs.CL","cs.CY"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-03-09T01:13:56Z","title_canon_sha256":"78e7f8bc44dd37e6b303c5569cc1e76b709c498f708e4f6b3cb80516079f55d0"},"schema_version":"1.0","source":{"id":"2503.06378","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06378","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06378v2","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06378","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"pith_short_12","alias_value":"73XVJFIDPLE4","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"pith_short_16","alias_value":"73XVJFIDPLE4YCRT","created_at":"2026-07-05T10:31:56Z"},{"alias_kind":"pith_short_8","alias_value":"73XVJFID","created_at":"2026-07-05T10:31:56Z"}],"graph_snapshots":[{"event_id":"sha256:3ed88b9699bb700928f1ed7a560e5ca301d83723c0f6fbcda2f2e36176780c4a","target":"graph","created_at":"2026-07-05T10:31:56Z","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/2503.06378/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Ensuring safe and effective use of AI requires understanding and anticipating its performance on novel tasks, from advanced scientific challenges to transformed workplace activities. So far, benchmarking has guided progress in AI, but it has offered limited explanatory and predictive power for general-purpose AI systems, given the low transferability across diverse tasks. In this paper, we introduce general scales for AI evaluation that can explain what common AI benchmarks really measure, extract ability profiles of AI systems, and predict their performance for new task instances, in- and out","authors_text":"Behzad Mehrbakhsh, David Stillwell, Fernando Mart\\'inez-Plumed, Jindong Wang, Jiyun Zu, John Burden, Jonathan E. Prunty, Jos\\'e Hern\\'andez-Orallo, Katherine M. Collins, Kexin Jiang Chen, Lexin Zhou, Lorenzo Pacchiardi, Lucy Cheke, Luning Sun, Manuel Cebrian, Pablo A. M. Casares, Pablo S\\'anchez-Garc\\'ia, Patrick C. Kyllonen, Peter Henderson, Qinlin Zhao, Seraphina Zhang, Sherry Tongshuang Wu, Xing Xie, Yael Moros-Daval, Yitian Huang, Zongqian Li","cross_cats":["cs.CL","cs.CY"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-03-09T01:13:56Z","title":"General Scales Unlock AI Evaluation with Explanatory and Predictive Power"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06378","kind":"arxiv","version":2},"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":"sha256:69ebfe62522c56f198b5c80f5a5e58a880624732f45a1762b24e7bc486d5aff3","target":"record","created_at":"2026-07-05T10:31:56Z","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":"10b86c7e748898bb949194fde48700de87f56553e85d6f0ada7af2ec2df82e41","cross_cats_sorted":["cs.CL","cs.CY"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-03-09T01:13:56Z","title_canon_sha256":"78e7f8bc44dd37e6b303c5569cc1e76b709c498f708e4f6b3cb80516079f55d0"},"schema_version":"1.0","source":{"id":"2503.06378","kind":"arxiv","version":2}},"canonical_sha256":"feef5495037ac9cc0a33bb2fd72439de02db98d5617f48129402aa2ce87e7d48","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"feef5495037ac9cc0a33bb2fd72439de02db98d5617f48129402aa2ce87e7d48","first_computed_at":"2026-07-05T10:31:56.243015Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:31:56.243015Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JNvD1TvafoXQx90dfK/GLkgpDbNcb5ydxbvkIZQr96jdvTVdXj8B28aBiYewmkbMtfuntIovVIcxT3dVgJwuAg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:31:56.243744Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.06378","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:69ebfe62522c56f198b5c80f5a5e58a880624732f45a1762b24e7bc486d5aff3","sha256:3ed88b9699bb700928f1ed7a560e5ca301d83723c0f6fbcda2f2e36176780c4a"],"state_sha256":"dabcb4448b188ac40ac5c36118cae95ded42eff400245dfbd15b29bbe99a579b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MMuhiNwzNT4bZsHa2+wZqUA0wQOIzb6X32uau1oJU+XrBU0uJsNYw/38ecvVcJh/dVBBmYuuCjABHQZ/S4o4Aw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T00:39:25.157804Z","bundle_sha256":"a1531736ad231ae48f807d4427efbb3b89311d8c8c869b3c782e3fce3be36e0f"}}