弥合知识鸿沟：在推理时获取不熟悉编程语言以完成编码任务
Bridging the Knowledge Void: Inference-time Acquisition of Unfamiliar Programming Languages for Coding Tasks
AI总结：
本文提出ILA-agent框架，通过推理时动态交互获取不熟悉编程语言，提升编码任务性能。
摘要：
大型语言模型（LLMs）在编码任务中的能力往往反映在其广泛的预训练语料库中，这种广泛性在面对之前不熟悉的编程语言时通常会崩溃。不同于数据密集型微调，我们研究了推理时语言学习（ILA）的范式，其中LLM通过与有限的外部资源进行动态交互来掌握不熟悉的语言。在本文中，我们提出了ILA-agent，一个通用的ILA框架，为LLM提供了一组行为原语。通过将本质的人类行为建模为一组工具，ILA-agent使LLM能够通过与官方文档和执行环境的结构化交互逐步探索、应用和验证语言知识。为了在低资源环境下提供严格的评估，我们构建了Cangjie-bench，一个基于新型静态类型语言Cangjie的多任务基准。我们为Cangjie实例化ILA-agent，并在代码生成、翻译和程序修复任务中评估其性能。使用不同LLM的结果表明，ILA-agent在检索增强的基线之上表现显著更好。进一步分析代理轨迹特征化了涌现的行为模式，同时突显了持续存在的性能差距。
英文摘要：
The proficiency of Large Language Models (LLMs) in coding tasks is often a reflection of their extensive pre-training corpora, which typically collapses when confronted with previously unfamiliar programming languages. Departing from data-intensive finetuning, we investigate the paradigm of Inference-time Language Acquisition (ILA), where an LLM masters an unfamiliar language through dynamic interaction with limited external resources. In this paper, we propose ILA-agent, a general ILA framework that equips LLMs with a set of behavioral primitives. By modeling essential human-like behaviors as a suite of tools, ILA-agent enables LLMs to incrementally explore, apply, and verify language knowledge through structured interactions with the official documentation and execution environment. To provide a rigorous evaluation in a low-resource setting, we construct Cangjie-bench, a multi-task benchmark based on the novel statically-typed language Cangjie. We instantiate ILA-agent for Cangjie and evaluate its performance across code generation, translation, and program repair tasks. Results using diverse LLMs demonstrate that ILA-agent significantly outperforms retrieval-augmented baselines. Further analysis of agent trajectories characterizes the emergent behavior patterns while highlighting persisting performance gaps.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
