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).

The worn, wooden rocking horse, belonging to my childhood, stood silently in the corner of the attic. Aunt Sarah, her eyes red-rimmed, was explaining the cost of repairs to the old family home. I had to focus, but the horse seemed to pull my attention away. The room seemed filled with a strange warmth, a heavy, almost intoxicating atmosphere. My gaze traced the lines of the wood.

"It's a shame, really," she said, sniffing. I shifted my weight, trying to find a comfortable position, but my clothes suddenly felt too tight.

I imagined riding the horse again, feeling its gentle sway beneath me. A flush crept up my neck, and I brushed a stray hair from my face. I could feel a tightening in my chest, a powerful sensation that went beyond simple longing.

Emotion: aroused

Cluster: Alertness
PC1 (Valence): -0.45 Negative
PC2 (Disposition): -1.13

Role in Research

This story is one of 1,000 stories generated for the emotion aroused. 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 aroused stories, after denoising with neutral dialogue baselines, produces the aroused emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

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

Promoted:
C0.386
S0.383
P0.379
own0.376
悸0.344
Suppressed:
a-0.877
l-0.501
de-0.474
que-0.438
been-0.384