{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2NEVFU4PLCEMUABHYGYFYZRB75","short_pith_number":"pith:2NEVFU4P","schema_version":"1.0","canonical_sha256":"d34952d38f5888ca0027c1b05c6621ff40da2cf94e6c7e4af92432baa0b25abb","source":{"kind":"arxiv","id":"2408.14354","version":1},"attestation_state":"computed","paper":{"title":"SWE-bench-java: A GitHub Issue Resolving Benchmark for Java","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.SE","authors_text":"Ailun Yu, Bei Guan, Bo Shen, Daoguang Zan, Dezhi Ran, Dong Chen, Guangtai Liang, Hao Yu, Lei Yu, Muhan Zeng, Pan Bian, Pengjie Huang, Qianxiang Wang, Shaoxin Lin, Tao Xie, Wei Liu, Yifan Shi, Yongji Wang, Zhirong Huang, Zongshuai Qi","submitted_at":"2024-08-26T15:30:05Z","abstract_excerpt":"GitHub issue resolving is a critical task in software engineering, recently gaining significant attention in both industry and academia. Within this task, SWE-bench has been released to evaluate issue resolving capabilities of large language models (LLMs), but has so far only focused on Python version. However, supporting more programming languages is also important, as there is a strong demand in industry. As a first step toward multilingual support, we have developed a Java version of SWE-bench, called SWE-bench-java. We have publicly released the dataset, along with the corresponding Docker"},"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":"2408.14354","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2024-08-26T15:30:05Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"6c20cf476f0f719b0c915045ae15ca217541bd1a20e2f9c9592e427317ac0ebf","abstract_canon_sha256":"ff7f619075d800b404e44f8ffba4e4822f2581d60212d906e43f97ffd07754dc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:59:19.649395Z","signature_b64":"kSp9YPaZmC42gRZE7C9efokQPrr0OtrY5Smc77ZRoFyeAT5FY2Vl8bcyJ/MbRcaFOSIXL4LboikKiwFVx0kOBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d34952d38f5888ca0027c1b05c6621ff40da2cf94e6c7e4af92432baa0b25abb","last_reissued_at":"2026-07-05T08:59:19.648833Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:59:19.648833Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SWE-bench-java: A GitHub Issue Resolving Benchmark for Java","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.SE","authors_text":"Ailun Yu, Bei Guan, Bo Shen, Daoguang Zan, Dezhi Ran, Dong Chen, Guangtai Liang, Hao Yu, Lei Yu, Muhan Zeng, Pan Bian, Pengjie Huang, Qianxiang Wang, Shaoxin Lin, Tao Xie, Wei Liu, Yifan Shi, Yongji Wang, Zhirong Huang, Zongshuai Qi","submitted_at":"2024-08-26T15:30:05Z","abstract_excerpt":"GitHub issue resolving is a critical task in software engineering, recently gaining significant attention in both industry and academia. Within this task, SWE-bench has been released to evaluate issue resolving capabilities of large language models (LLMs), but has so far only focused on Python version. However, supporting more programming languages is also important, as there is a strong demand in industry. As a first step toward multilingual support, we have developed a Java version of SWE-bench, called SWE-bench-java. We have publicly released the dataset, along with the corresponding Docker"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.14354","kind":"arxiv","version":1},"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/2408.14354/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":"2408.14354","created_at":"2026-07-05T08:59:19.648898+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.14354v1","created_at":"2026-07-05T08:59:19.648898+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.14354","created_at":"2026-07-05T08:59:19.648898+00:00"},{"alias_kind":"pith_short_12","alias_value":"2NEVFU4PLCEM","created_at":"2026-07-05T08:59:19.648898+00:00"},{"alias_kind":"pith_short_16","alias_value":"2NEVFU4PLCEMUABH","created_at":"2026-07-05T08:59:19.648898+00:00"},{"alias_kind":"pith_short_8","alias_value":"2NEVFU4P","created_at":"2026-07-05T08:59:19.648898+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26979","citing_title":"How Much Static Structure Do Code Agents Need? A Study of Deterministic Anchoring","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26979","citing_title":"How Much Static Structure Do Code Agents Need? A Study of Deterministic Anchoring","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08840","citing_title":"Beyond Pass Rate: A Multilingual, Execution-Grounded Evaluation of Open Code LLMs","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29709","citing_title":"Bash-Commenter: Leveraging Syntax-Aware Preference Optimization to Reinforce Large Language Model for Bash Code Comment Generation","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25285","citing_title":"PR-Aware Automated Unit Test Generation: Challenges and Opportunities","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2409.02977","citing_title":"Large Language Model-Based Agents for Software Engineering: A Survey","ref_index":251,"is_internal_anchor":false},{"citing_arxiv_id":"2604.00824","citing_title":"Yet Even Less Is Even Better For Agentic, Reasoning, and Coding LLMs","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75","json":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75.json","graph_json":"https://pith.science/api/pith-number/2NEVFU4PLCEMUABHYGYFYZRB75/graph.json","events_json":"https://pith.science/api/pith-number/2NEVFU4PLCEMUABHYGYFYZRB75/events.json","paper":"https://pith.science/paper/2NEVFU4P"},"agent_actions":{"view_html":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75","download_json":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75.json","view_paper":"https://pith.science/paper/2NEVFU4P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.14354&json=true","fetch_graph":"https://pith.science/api/pith-number/2NEVFU4PLCEMUABHYGYFYZRB75/graph.json","fetch_events":"https://pith.science/api/pith-number/2NEVFU4PLCEMUABHYGYFYZRB75/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75/action/storage_attestation","attest_author":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75/action/author_attestation","sign_citation":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75/action/citation_signature","submit_replication":"https://pith.science/pith/2NEVFU4PLCEMUABHYGYFYZRB75/action/replication_record"}},"created_at":"2026-07-05T08:59:19.648898+00:00","updated_at":"2026-07-05T08:59:19.648898+00:00"}