Consciousness Benchmark
We're an independent research effort in SF designing evals for AI consciousness (ConsciousBench).
Our first evals will be a set of 100 questions representing each major branch of philosophy. We will test both closed and open-source models on this benchmark to attain a consciousness score from each philosophical perspective.
In the future, we will design task-oriented and spacial-awareness evals, along with others. If you're interested to contribute, please dm Koko at @kokoxsu.