These are images generated by a VQGAN-CLIP neural network (using the imagenet 16384 model), when the network was fed the corresponding text-based prompt shown below each image set. The images it produced are ordered by increasing depth, from left to right.
The network was created by Katherine Crowson.
"Hatred"
"Love"
"Rage"
"Anxiety"
"Boredom"
"Grief"
"Wrath"
"Schizophrenia"
"Rapture"
"Chaos"
"Psychosis"
"Depression"
"Paradise"
"Dreams"
"Cosmic"
"Divine"
"Eternity"
"Lust"
"Sorrow"