Similarity judgment

Same / Different

reconstruct human perceptual spaces; benchmark AI against them.

AI safety

Cognitive task

Captured cognition human similarity spaces plus the "is the model's space human-like?" benchmark the field badly needs

Local atlas

Cognition dashboard

0signals
0public
0avg confidence
0msmedian latency
Submit a signal to see the first reflection.

Commons loop

Recent signals

Cloudflare deployment

ordinal embedding / MDS to reconstruct the human perceptual space