BibbyResearch/climate-action-gpt-yale

ClimateAction-GPT-Yale 🌍

ClimateAction-GPT-Yale is a specialized large language model fine-tuned from mistralai/Mistral-7B-Instruct-v0.2. Developed at Yale University in collaboration with Aza Allsop and Leor Roseman, this model is designed to understand and generate text related to climate change, with a specific focus on community-level climate action and policy.

It was trained on the - Reworlding dataset by Leor Roseman Exeter university and Clima500 dataset, a benchmark collection for climate change-related misinformation detection, making it particularly adept at navigating complex and often contentious climate narratives.

🚀 Model Description

This model is the result of a research project at the Yale School of the Environment aimed at creating AI tools to support climate communicators, policymakers, and community organizers. By fine-tuning the powerful Mistral-7B-Instruct model on a specialized dataset, ClimateAction-GPT-Yale can assist with tasks such as:

• Summarizing climate action plans.
• Answering questions about local climate initiatives.
• Drafting communications for community outreach.
• Identifying and analyzing claims related to climate change.

🔧 Intended Use

This model is intended for researchers, students, and practitioners in the climate space. It can be used as a tool for text generation, analysis, and as a conversational agent.

Example Use Cases:

• A nonprofit could use it to draft grant proposals for local green infrastructure projects.
• A student could use it to understand the key arguments in a municipal climate policy document.
• A researcher could use it to analyze public comments on proposed environmental regulations.

# Example of how to use the model with the transformers library
from transformers import pipeline

# Make sure to replace "your-username/your-model-name" with the actual model repo
pipe = pipeline("text-generation", model="your-username/your-model-name")

prompt = """
[INST] As a climate action assistant, please explain three practical steps a small coastal town can take to improve its resilience to rising sea levels. [/INST]
"""

result = pipe(prompt, max_new_tokens=250)
print(result[0]['generated_text'])