知识条件下的可执行Unity游戏场景单遍大语言模型合成：26个目标可玩概念的编译器错误普查
Knowledge-Conditioned, Single-Pass LLM Synthesis of Executable Unity Game Scenes: A Compiler Error Census across 26 Goal Playable Concepts
浏览论文内容
中文总结 AI 辅助
研究大语言模型为Unity游戏场景编写代码，去除迭代修复循环进行单遍生成评估，通过对多个模型、模式等生成的代码分析编译器错误，发现瓶颈是缺少引擎特定知识，按需求对目标模式排序展示单遍生成断点。
中文摘要 AI 辅助
大语言模型（LLMs）为游戏场景编写Unity C#代码。然而，几乎所有演示都依赖于迭代修复循环，即不断重新生成代码直到编译成功，这混淆了模型编写的内容和循环修复的内容。我们去除了该循环并评估单遍生成，即初稿即为最终稿。这隔离了模型的参数知识，这是对无辅助生成的最严格测试。模型针对10400次生成实例化了目标可玩概念（目标模式的可玩对应物）（四个开放权重模型，7B - 30B；两种生成模式；四个中间表示（IR）条件级别；26个目标模式；20个种子）。没有一个编译成可运行的场景，不存在生存偏差。为了解生成的C#脚本如何失败，我们将90673次编译器错误发生背后的99个错误代码分类为基础错误（发明或误用的Unity类型和API）或卫生错误（无需Unity知识的结构缺陷）。不同目标模式的分类差异很大（例如，隐身模式大多在发明的引擎引用上失败；捕获模式在普通C#结构上失败）。更大的模型、更严格的IR和不同的生成模式会改变错误，但从未产生可编译的场景。瓶颈在于缺少特定于引擎的知识。普查按该需求对目标模式进行排序，向设计师展示单遍生成在何处中断。
英文摘要
Large language models (LLMs) write Unity C\# for game scenes. Yet nearly all demonstrations rest on an iterative repair loop that regenerates code until it compiles, conflating what the model writes with what the loop fixes. We remove the loop and evaluate a single pass, where the first draft is final. This isolates the model's parametric knowledge, the most stringent test of unaided generation. Models instantiate Goal Playable Concepts, playable counterparts of goal patterns, across 10,400 generations (four open-weight models, 7B--30B; two generation modes; four intermediate-representation (IR) conditioning levels; 26 goal patterns; 20 seeds). None compiled into a runnable scene, leaving no survivorship bias. To understand how the generated C\# scripts fail, we categorize the 99 error codes behind 90{,}673 compiler-error occurrences as Grounding (invented or misused Unity types and APIs) or Hygiene (structural defects needing no Unity knowledge). The split differs sharply by goal pattern (e.g., Stealth fails mostly on invented engine references; Capture on plain C\# structure). Larger models, stricter IRs, and different generation modes move the errors but never yield a compiling scene. The bottleneck is missing engine-specific knowledge. The census orders goal patterns by that demand, showing designers where single-pass generation breaks.
发表机构
- Chalmers University of Technology（查尔姆斯理工大学）
- University of Gothenburg（哥德堡大学）
机构由 AI 辅助整理，请以论文原文为准。
