About This Research

Investigating AI's Understanding of Dog Breeds

Research Question

What we're investigating

How well do large language models understand dog breeds and their characteristics? This project aims to validate AI-generated trait associations against real-world experiences from dog owners and enthusiasts.

By collecting votes from the community, we can measure the alignment between AI predictions and human knowledge, potentially revealing both the strengths and limitations of AI in understanding complex breed-trait relationships.

Methodology

How the data is collected and analyzed

The AI scores are generated using large language models that analyze breed descriptions and characteristics. Each breed-trait pair receives a similarity score based on the model's understanding.

Community votes are collected anonymously and stored on a Firebase database. This lightweight approach allows us to gather feedback without requiring user accounts or maintaining a heavy database infrastructure. All of the data is freely available for research purposes.

Contribute

How you can help

  • Vote on breed-trait associations to validate or challenge AI predictions
  • Share your expertise if you have experience with specific breeds
  • Explore how AI understands different aspects of dog breeds
  • Help identify where AI might be making incorrect assumptions by voting down and highlight the ones you see as correct by voting up.

Data & Privacy

How we handle information

We don't track any personal information or require any login. The project data is freely available for research purposes for anyone that wants to use it. Further research is subject to getting a meaningful amount of votes.