{
  "added_at": "2026-11-10",
  "arxiv_id": "2610.28212",
  "arxiv_version_reviewed": "v1",
  "authors": [
    "Akarsh Ashok",
    "Dinil Mon Divakaran",
    "Vivek Balachandran"
  ],
  "bibtex_url": "https://awesomerea.org/papers/citations/paper-exebench-fhd.bib",
  "code_url": null,
  "date_basis": "First arXiv submission date.",
  "date_basis_zh": "arXiv 首次提交日期。",
  "date_precision": "day",
  "description_en": "The paper identifies nine fidelity issues in ExeBench and introduces ExeBench-FHD, adding assembly, data, wrapper and warning fields to improve LLM decompiler evaluation.",
  "description_zh": "论文指出 ExeBench 中九项影响评测保真度的问题，并提出 ExeBench-FHD，加入汇编、数据段、测试包装器和警告标记等字段，以改进大语言模型反编译器的评测。",
  "doi": null,
  "entry_url": "https://awesomerea.org/papers/paper-exebench-fhd/",
  "entry_url_zh": "https://awesomerea.org/zh/papers/paper-exebench-fhd/",
  "evidence_urls": [
    "https://arxiv.org/abs/2610.28212"
  ],
  "first_arxiv_date": "2026-10-20",
  "id": "paper-exebench-fhd",
  "json_url": "https://awesomerea.org/papers/paper-exebench-fhd/record.json",
  "kind": "research",
  "markdown_url": "https://awesomerea.org/papers/paper-exebench-fhd/index.md",
  "markdown_url_zh": "https://awesomerea.org/zh/papers/paper-exebench-fhd/index.md",
  "metadata_notes": "The date and venue come from the arXiv record.",
  "metadata_notes_zh": "日期和发表场所来自 arXiv 记录。",
  "pdf_url": "https://arxiv.org/pdf/2610.28212",
  "project_url": null,
  "publication_date": "2026-10-20",
  "publication_status": "preprint",
  "record_type": "paper",
  "runtime_tested": false,
  "scope_limitations_en": "The evaluation measures GPT 6 Sol's re-execution rate at -O2 on four architectures, including functions that read data sections, mix parameter types or have unused parameters.",
  "scope_limitations_zh": "评测在四种架构的 -O2 编译结果上测量 GPT 6 Sol 的重新执行成功率，涵盖读取数据段、混用参数类型或含有未使用参数的函数。",
  "targets": [
    "binary"
  ],
  "title": "ExeBench-FHD: Addressing Fidelity Issues in Decompiler Evaluation",
  "title_zh": "ExeBench-FHD：解决反编译器评测中的保真度问题",
  "url": "https://arxiv.org/abs/2610.28212",
  "venue": "arXiv",
  "verification": "source-reviewed",
  "verified_at": "2026-11-09",
  "why_fits_en": "It analyzes and improves a benchmark for recovering C code from compiled functions.",
  "why_fits_zh": "它分析并改进了用于从已编译函数恢复 C 代码的评测基准。",
  "year": 2026
}
