"Dataset of Prompting Techniques Evaluation for LLM Agents: Prompts, Scores, and Code Across Five Test Scenarios"
收藏DataCite Commons2025-09-04 更新2026-05-03 收录
下载链接：
https://ieee-dataport.org/documents/dataset-prompting-techniques-evaluation-llm-agents-prompts-scores-and-code-across-five
下载链接
链接失效反馈官方服务：
资源简介：
"This dataset is associated with the applied research article titled \u201cEvaluating the Effectiveness of Prompting Techniques for LLMs on Agent-Based Tasks Requiring Complex Conditional Reasoning.\u201dIt comprises three primary components:The complete set of prompt inputs used in all five test scenarios.Comprehensive evaluation results for eight prompting techniques, assessed across five test scenarios and four large language models (LLMs).The source code used to run and reproduce all experiments.The dataset is intended to support reproducibility and facilitate further research into prompting strategies for LLM-based agents operating in tasks that require multi-step reasoning and conditional decision-making."
提供机构：
IEEE DataPort
创建时间：
2025-09-04
