{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:B7UBF3MUMIAGE52UHF6RUEBFIN","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":"6b750a332032655ea70fbe6907d831abf605223082d94c705936b23ba5f3c60b","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-01-30T16:56:08Z","title_canon_sha256":"cabc85aa95577424e3da2eacaf190de19f9f470f6f0cc64b7c429697d6a3294b"},"schema_version":"1.0","source":{"id":"2501.18482","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.18482","created_at":"2026-07-05T10:07:34Z"},{"alias_kind":"arxiv_version","alias_value":"2501.18482v1","created_at":"2026-07-05T10:07:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.18482","created_at":"2026-07-05T10:07:34Z"},{"alias_kind":"pith_short_12","alias_value":"B7UBF3MUMIAG","created_at":"2026-07-05T10:07:34Z"},{"alias_kind":"pith_short_16","alias_value":"B7UBF3MUMIAGE52U","created_at":"2026-07-05T10:07:34Z"},{"alias_kind":"pith_short_8","alias_value":"B7UBF3MU","created_at":"2026-07-05T10:07:34Z"}],"graph_snapshots":[{"event_id":"sha256:8049a2e08bf56169dcf3eae4a3b621e15c2ae7b8b701e8e227db050bc9daf390","target":"graph","created_at":"2026-07-05T10:07:34Z","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/2501.18482/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Code Executing Reasoning is becoming a new non-functional metric that assesses the ability of large language models (LLMs) in programming tasks. State-of-the-art frameworks (CodeMind or REval) and benchmarks (CruxEval) usually focus on LLM's prediction of a given code's input/output or intermediate variable states/values on limited programs. However, there is no tool for more in-depth analysis of the results. Without such a tool, the observations about LLM's code execution reasoning cannot be generalized to more datasets, preventing the research community and practitioners from devising the ne","authors_text":"Changshu Liu, Reyhaneh Jabbarvand","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-01-30T16:56:08Z","title":"A Tool for In-depth Analysis of Code Execution Reasoning of Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.18482","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":"sha256:3af4b0dd5b485711b0023548fec1724bf1f5fad067bb0337a77d4caeb2acbf78","target":"record","created_at":"2026-07-05T10:07:34Z","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":"6b750a332032655ea70fbe6907d831abf605223082d94c705936b23ba5f3c60b","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-01-30T16:56:08Z","title_canon_sha256":"cabc85aa95577424e3da2eacaf190de19f9f470f6f0cc64b7c429697d6a3294b"},"schema_version":"1.0","source":{"id":"2501.18482","kind":"arxiv","version":1}},"canonical_sha256":"0fe812ed946200627754397d1a1025436d6cb530f03b640e4323671cfffc4375","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0fe812ed946200627754397d1a1025436d6cb530f03b640e4323671cfffc4375","first_computed_at":"2026-07-05T10:07:34.636032Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:07:34.636032Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"2AA4pBghRBVXKIRWRzt9kwx6c6W1KmpRAXZS322JS+KnWbi+xNJwQmWjqAkfJD+fOED8RgVYsWA0tcwCHj+0AQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:07:34.636474Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.18482","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3af4b0dd5b485711b0023548fec1724bf1f5fad067bb0337a77d4caeb2acbf78","sha256:8049a2e08bf56169dcf3eae4a3b621e15c2ae7b8b701e8e227db050bc9daf390"],"state_sha256":"f6bae6d3812781aba43b3d72b8254aee9a458b973cb096e0db5ffe831aad6021"}