Computational experiments within metaverse service ecosystems enable the identification of social risks and governance crises, and the optimization of governance strategies through counterfactual inference to dynamically guide real-world service ecosystem operations. The advent of Large Language Models (LLMs) has empowered LLM-based agents to function as autonomous service entities capable of executing diverse service operations within metaverse ecosystems, thereby facilitating the governance of metaverse service ecosystem with computational experiments.

Code search is essential for code reuse, allowing developers to efficiently locate relevant code snippets. Traditional encoder-based models, however, face challenges with poor generalization and input length limitations. In contrast, decoder-only large language models (LLMs), with their larger size, extensive pre-training, and ability to handle longer inputs, present a promising solution to these issues. However, their effectiveness in code search has not been fully explored.

CodePromptEval is a dataset of 7072 prompts designed to evaluate five prompt techniques (few-shot, persona, chain-of-thought, function signature, list of packages) and their effect on the correctness, similarity, and quality of complete functions generated. Each data point in the dataset includes a function generation task, a combination of prompt techniques to be applied, the prompt in natural language that applied the prompt techniques, the ground truth of the functions (human-written functions based on CoderEval dataset by Yu et al.), the tests to evaluate the correctness of the generate