LLM Preference Alignment: Stakeholder Data for Chinese Urban Planning

Shimin Qi produced this dataset in 2026 to support Reinforcement Learning from Human Feedback (RLHF) for large language models in urban planning. It comprises raw redevelopment data from Chinese municipal websites and multi-stakeholder annotated preference pairs used to fine-tune ChatGLM3-6B.

Use Cases

• Fine-tuning LLMs using the multi-stakeholder annotated preference data to simulate public participation
• Analyzing urban redevelopment cases collected from Chinese municipal websites for policy research
• Benchmarking ChatGLM3-6B using the provided training and analysis scripts

Strengths

• Multi-stakeholder annotated preference data
• Primary source data from official Chinese municipal websites
• Includes training and analysis scripts for ChatGLM3-6B

Limitations

• Geographic bias limited to major Chinese cities
• Specific to urban redevelopment domain
• Unknown record count and file structure

Provenance

Source

Official websites of major Chinese cities

Collection Method

Scraped from municipal websites and annotated by stakeholders

Freshness

Last updated March 2026

Geography

China