About HypothesisAI
HypothesisAI is a platform for evaluating how well AI models generate scientific hypotheses. Multiple AI models from different providers generate hypotheses across scientific domains, and domain experts rate them on novelty, plausibility, and testability. The results feed into a leaderboard that ranks models by their average expert ratings.
How It Works
- Sign in with your Google account
- Pick a scientific domain you have expertise in
- Read an AI-generated hypothesis and rate it on three criteria (1-5 scale)
- Your ratings are aggregated with other experts to rank the AI models
Rating Criteria
Novelty — How original is the hypothesis compared to existing theories?
Plausibility — How well-supported is it by existing scientific knowledge?
Testability — How feasible is it to validate through experiments?