Player Level Completion and Ratings

该数据集汇总了使用传统程序内容生成（PCG）以及大型语言模型（LLM）生成的关卡中，玩家的完成率和评分信息。其中数据包括了LLM生成的关卡完成率（55%）与传统关卡（35%）的完成率，以及玩家对LLM关卡（平均评分为3.87/5）与传统PCG关卡（平均评分为4.22/5）的平均评分。该任务旨在评估个性化关卡生成与传统关卡生成在保持玩家参与度和提升玩家满意度方面的有效性。This dataset aggregates player completion rates and rating metrics across game levels generated via two approaches: traditional procedural content generation (PCG) and large language models (LLMs). It contains the completion rates of LLM-generated levels (55%) and traditionally PCG-generated levels (35%), alongside the average player ratings for LLM-generated levels (mean score: 3.87/5) and traditional PCG levels (mean score: 4.22/5). The objective of this dataset is to assess the efficacy of personalized level generation powered by LLMs relative to traditional level generation in sustaining player engagement and enhancing player satisfaction.

搜集汇总

背景与挑战

背景概述

该数据集聚焦于玩家对传统PCG与LLM生成关卡的完成率和评分对比，数据显示LLM关卡完成率更高（55% vs 35%），但传统关卡平均评分略优（4.22/5 vs 3.87/5），旨在评估两者在玩家参与度和满意度方面的效果差异，为个性化关卡生成研究提供数据支持。