{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:WNYDD5L53VSGMALCHABJVIWPVN","short_pith_number":"pith:WNYDD5L5","canonical_record":{"source":{"id":"2402.17396","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T10:44:52Z","cross_cats_sorted":["cs.AI","cs.NE"],"title_canon_sha256":"6569d43e43c01f05e5d07d37fda1015bd97f8bfcb7ce4f9e04fc94da43eaf473","abstract_canon_sha256":"43962ed8f4c058a4fe93bc3245c0584b72bfd4d4de36100d8ab6dfb4a5b01cc2"},"schema_version":"1.0"},"canonical_sha256":"b37031f57ddd6466016238029aa2cfab6426a87ecb58e3ab12c8fdee19413457","source":{"kind":"arxiv","id":"2402.17396","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.17396","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"arxiv_version","alias_value":"2402.17396v2","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.17396","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"pith_short_12","alias_value":"WNYDD5L53VSG","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"pith_short_16","alias_value":"WNYDD5L53VSGMALC","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"pith_short_8","alias_value":"WNYDD5L5","created_at":"2026-07-05T08:42:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:WNYDD5L53VSGMALCHABJVIWPVN","target":"record","payload":{"canonical_record":{"source":{"id":"2402.17396","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T10:44:52Z","cross_cats_sorted":["cs.AI","cs.NE"],"title_canon_sha256":"6569d43e43c01f05e5d07d37fda1015bd97f8bfcb7ce4f9e04fc94da43eaf473","abstract_canon_sha256":"43962ed8f4c058a4fe93bc3245c0584b72bfd4d4de36100d8ab6dfb4a5b01cc2"},"schema_version":"1.0"},"canonical_sha256":"b37031f57ddd6466016238029aa2cfab6426a87ecb58e3ab12c8fdee19413457","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:42:33.392366Z","signature_b64":"osGIiG2gdZiyXy8bX6mq61+6zY4FmtBoyVT5pkgIVKYUeWd/RGqjqb884o2E6zI8vLSFcuzul/sjjN/x5gHWCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b37031f57ddd6466016238029aa2cfab6426a87ecb58e3ab12c8fdee19413457","last_reissued_at":"2026-07-05T08:42:33.391818Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:42:33.391818Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.17396","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":"2026-07-05T08:42:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"g0kdR9x2kzjLkJfKfuYSRoJALnEdo4FF/nxb0eiVffFMNM8hZ48ysbw4fFCDe4ygAueLyvzIsHHOFxOYmkX7Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T06:15:58.105826Z"},"content_sha256":"f91cffc3c95e174e39157a89ae9a6bd10e53ed2a402761380c329d7b06958603","schema_version":"1.0","event_id":"sha256:f91cffc3c95e174e39157a89ae9a6bd10e53ed2a402761380c329d7b06958603"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:WNYDD5L53VSGMALCHABJVIWPVN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Benchmarking GPT-4 on Algorithmic Problems: A Systematic Evaluation of Prompting Strategies","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.NE"],"primary_cat":"cs.CL","authors_text":"Alberto Testolin, Alessandro Sperduti, Flavio Petruzzellis","submitted_at":"2024-02-27T10:44:52Z","abstract_excerpt":"Large Language Models (LLMs) have revolutionized the field of Natural Language Processing thanks to their ability to reuse knowledge acquired on massive text corpora on a wide variety of downstream tasks, with minimal (if any) tuning steps. At the same time, it has been repeatedly shown that LLMs lack systematic generalization, which allows to extrapolate the learned statistical regularities outside the training distribution. In this work, we offer a systematic benchmarking of GPT-4, one of the most advanced LLMs available, on three algorithmic tasks characterized by the possibility to control"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.17396","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/2402.17396/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":"2026-07-05T08:42:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zeyg3OvBDxhOccQQ5OaRyXVDHlWTDfRX0FcV2GKi7YCGJBI2uzKwCWi+P+7Ev+nfkGzqqCr7MNzaX+r5iOXfCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T06:15:58.106389Z"},"content_sha256":"03d1bd4f22f35822569eba0e24c40b55866234d9395c6f96f16907b858b433d0","schema_version":"1.0","event_id":"sha256:03d1bd4f22f35822569eba0e24c40b55866234d9395c6f96f16907b858b433d0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WNYDD5L53VSGMALCHABJVIWPVN/bundle.json","state_url":"https://pith.science/pith/WNYDD5L53VSGMALCHABJVIWPVN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WNYDD5L53VSGMALCHABJVIWPVN/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":"2026-08-17T06:15:58Z","links":{"resolver":"https://pith.science/pith/WNYDD5L53VSGMALCHABJVIWPVN","bundle":"https://pith.science/pith/WNYDD5L53VSGMALCHABJVIWPVN/bundle.json","state":"https://pith.science/pith/WNYDD5L53VSGMALCHABJVIWPVN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WNYDD5L53VSGMALCHABJVIWPVN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:WNYDD5L53VSGMALCHABJVIWPVN","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":"43962ed8f4c058a4fe93bc3245c0584b72bfd4d4de36100d8ab6dfb4a5b01cc2","cross_cats_sorted":["cs.AI","cs.NE"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T10:44:52Z","title_canon_sha256":"6569d43e43c01f05e5d07d37fda1015bd97f8bfcb7ce4f9e04fc94da43eaf473"},"schema_version":"1.0","source":{"id":"2402.17396","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.17396","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"arxiv_version","alias_value":"2402.17396v2","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.17396","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"pith_short_12","alias_value":"WNYDD5L53VSG","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"pith_short_16","alias_value":"WNYDD5L53VSGMALC","created_at":"2026-07-05T08:42:33Z"},{"alias_kind":"pith_short_8","alias_value":"WNYDD5L5","created_at":"2026-07-05T08:42:33Z"}],"graph_snapshots":[{"event_id":"sha256:03d1bd4f22f35822569eba0e24c40b55866234d9395c6f96f16907b858b433d0","target":"graph","created_at":"2026-07-05T08:42:33Z","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/2402.17396/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have revolutionized the field of Natural Language Processing thanks to their ability to reuse knowledge acquired on massive text corpora on a wide variety of downstream tasks, with minimal (if any) tuning steps. At the same time, it has been repeatedly shown that LLMs lack systematic generalization, which allows to extrapolate the learned statistical regularities outside the training distribution. In this work, we offer a systematic benchmarking of GPT-4, one of the most advanced LLMs available, on three algorithmic tasks characterized by the possibility to control","authors_text":"Alberto Testolin, Alessandro Sperduti, Flavio Petruzzellis","cross_cats":["cs.AI","cs.NE"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T10:44:52Z","title":"Benchmarking GPT-4 on Algorithmic Problems: A Systematic Evaluation of Prompting Strategies"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.17396","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:f91cffc3c95e174e39157a89ae9a6bd10e53ed2a402761380c329d7b06958603","target":"record","created_at":"2026-07-05T08:42:33Z","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":"43962ed8f4c058a4fe93bc3245c0584b72bfd4d4de36100d8ab6dfb4a5b01cc2","cross_cats_sorted":["cs.AI","cs.NE"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T10:44:52Z","title_canon_sha256":"6569d43e43c01f05e5d07d37fda1015bd97f8bfcb7ce4f9e04fc94da43eaf473"},"schema_version":"1.0","source":{"id":"2402.17396","kind":"arxiv","version":2}},"canonical_sha256":"b37031f57ddd6466016238029aa2cfab6426a87ecb58e3ab12c8fdee19413457","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b37031f57ddd6466016238029aa2cfab6426a87ecb58e3ab12c8fdee19413457","first_computed_at":"2026-07-05T08:42:33.391818Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:42:33.391818Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"osGIiG2gdZiyXy8bX6mq61+6zY4FmtBoyVT5pkgIVKYUeWd/RGqjqb884o2E6zI8vLSFcuzul/sjjN/x5gHWCg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:42:33.392366Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.17396","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f91cffc3c95e174e39157a89ae9a6bd10e53ed2a402761380c329d7b06958603","sha256:03d1bd4f22f35822569eba0e24c40b55866234d9395c6f96f16907b858b433d0"],"state_sha256":"dedbb8c7d37d3bcf9facf844c60ed7ed3439e49b2e6e93fee9e33d97c6db96c8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Sf7bVDzIhMY2WYum0lXxNoaIYLv4ZfjtpZ9RDpDMmbFEvMUIG+/xnCissyqDaBR+KXHK1mWgvHdHKj2pcMRMCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-17T06:15:58.109756Z","bundle_sha256":"47e342c5a7c568f6e39407d4719bbf63819a59fb67fbf2240c7095c25b503b3f"}}