{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ZADZ5FBWVXB2JBP2LL5GRQXBIM","short_pith_number":"pith:ZADZ5FBW","schema_version":"1.0","canonical_sha256":"c8079e9436adc3a485fa5afa68c2e143290c3e9416b6f80e11ba608229a624e8","source":{"kind":"arxiv","id":"2502.11620","version":3},"attestation_state":"computed","paper":{"title":"Assessing Correctness in LLM-Based Code Generation via Uncertainty Estimation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Arindam Sharma, Cristina David","submitted_at":"2025-02-17T10:03:01Z","abstract_excerpt":"In this work, we explore uncertainty estimation as a proxy for correctness in LLM-generated code. To this end, we adapt two state-of-the-art techniques from natural language generation -- one based on entropy and another on mutual information -- to the domain of code generation. Given the distinct semantic properties of code, we introduce modifications, including a semantic equivalence check based on symbolic execution. Our findings indicate a strong correlation between the uncertainty computed through these techniques and correctness, highlighting the potential of uncertainty estimation for q"},"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":"2502.11620","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-02-17T10:03:01Z","cross_cats_sorted":[],"title_canon_sha256":"33d3c1e09a9908e7ecd51e1a4ccb5f51e5654078f7268d59b2b713fc81478568","abstract_canon_sha256":"79fba88d00c01543f045dc07d9aa6b47a13a2a5f271c04cd6749ba863685a476"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:29:55.356175Z","signature_b64":"VUnF6tnYOdqfflnMX50/F8OC0uaJKZHXDds0g8bUkYHYc/HgWGM6vK1w+wl4uu/DkcPGNy+1nBJoz5twDFBqBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c8079e9436adc3a485fa5afa68c2e143290c3e9416b6f80e11ba608229a624e8","last_reissued_at":"2026-07-05T11:29:55.355690Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:29:55.355690Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Assessing Correctness in LLM-Based Code Generation via Uncertainty Estimation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Arindam Sharma, Cristina David","submitted_at":"2025-02-17T10:03:01Z","abstract_excerpt":"In this work, we explore uncertainty estimation as a proxy for correctness in LLM-generated code. To this end, we adapt two state-of-the-art techniques from natural language generation -- one based on entropy and another on mutual information -- to the domain of code generation. Given the distinct semantic properties of code, we introduce modifications, including a semantic equivalence check based on symbolic execution. Our findings indicate a strong correlation between the uncertainty computed through these techniques and correctness, highlighting the potential of uncertainty estimation for q"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.11620","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/2502.11620/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":"2502.11620","created_at":"2026-07-05T11:29:55.355751+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.11620v3","created_at":"2026-07-05T11:29:55.355751+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.11620","created_at":"2026-07-05T11:29:55.355751+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZADZ5FBWVXB2","created_at":"2026-07-05T11:29:55.355751+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZADZ5FBWVXB2JBP2","created_at":"2026-07-05T11:29:55.355751+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZADZ5FBW","created_at":"2026-07-05T11:29:55.355751+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.01953","citing_title":"Underspecification does not imply Incoherence: The Risks of Semantic Collapse in Coding Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01953","citing_title":"Underspecification does not imply Incoherence: The Risks of Semantic Collapse in Coding Models","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28500","citing_title":"Functional Entropy: Predicting Functional Correctness in LLM-Generated Code with Uncertainty Quantification","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2603.27098","citing_title":"Ensemble-Based Uncertainty Estimation for Code Correctness Estimation","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09023","citing_title":"Using Semantic Distance to Estimate Uncertainty in LLM-Based Code Generation","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM","json":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM.json","graph_json":"https://pith.science/api/pith-number/ZADZ5FBWVXB2JBP2LL5GRQXBIM/graph.json","events_json":"https://pith.science/api/pith-number/ZADZ5FBWVXB2JBP2LL5GRQXBIM/events.json","paper":"https://pith.science/paper/ZADZ5FBW"},"agent_actions":{"view_html":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM","download_json":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM.json","view_paper":"https://pith.science/paper/ZADZ5FBW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.11620&json=true","fetch_graph":"https://pith.science/api/pith-number/ZADZ5FBWVXB2JBP2LL5GRQXBIM/graph.json","fetch_events":"https://pith.science/api/pith-number/ZADZ5FBWVXB2JBP2LL5GRQXBIM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM/action/storage_attestation","attest_author":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM/action/author_attestation","sign_citation":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM/action/citation_signature","submit_replication":"https://pith.science/pith/ZADZ5FBWVXB2JBP2LL5GRQXBIM/action/replication_record"}},"created_at":"2026-07-05T11:29:55.355751+00:00","updated_at":"2026-07-05T11:29:55.355751+00:00"}