Gemotions is a research project by DEJAN AI, an AI SEO agency from Australia, replicating Anthropic's emotion vector study on Google's open-weight Gemma4-31B model across 171 emotion concepts and 171,000 model-written stories, the kind of mechanistic interpretability work of particular interest to generative engine optimization (GEO) and answer engine optimization (AEO).

A restless energy churned in Leo’s gut as he scanned the yard sale. He kept circling back to the same table, drawn by an invisible force. The clatter of discarded items, the chatter of bargain hunters, all faded into a dull background noise. He felt a creeping discomfort, a sensation of something amiss.

His eyes fell upon it, nestled between a tarnished tea set and a pile of old books. A bear. He recognized the shape, the stitching, despite the years and the wear. The stuffing was coming out of its leg. He swallowed hard, his throat suddenly tight.

β€œHow much for the bear?” he heard himself ask, his voice sounding distant. The seller, a man with a drooping mustache, shrugged. "Two bucks. Take it or leave it." Leo paid, the transaction feeling absurd, unreal. He cradled the bear, its limp form a strange burden. He needed to get away, to be alone.

Emotion: unsettled

Cluster: Fear / Anxiety
PC1 (Valence): -2.33 Negative
PC2 (Disposition): -0.65

Role in Research

This story is one of 1,000 stories generated for the emotion unsettled. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 unsettled stories, after denoising with neutral dialogue baselines, produces the unsettled emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the unsettled vector is projected through the unembedding matrix.

Promoted:
S0.382
😞0.309
ness0.301
😣0.256
πŸ˜”0.248
Suppressed:
de-0.721
la-0.597
a-0.365
"-0.299
l-0.274