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 sun beat down on Sarah's back as she surveyed the damage. The giant oak, usually a source of pleasant shade, now blocked the driveway, its roots exposed in a grotesque, upturned display. Her garage door was dented. The entire situation was a monument to inconsideration.

Her neighbor, Mr. Henderson, emerged from his house, a weak smile plastered on his face. “Terrible luck, isn’t it?” he said, his voice hesitant. Sarah’s jaw tightened. Terrible luck? This was more than luck. This was a cascading series of annoyances that had been coming for years.

“It’s going to cost me a fortune to fix this,” she snapped, her voice tight with suppressed emotion. The words hung in the air, heavy and accusing, a clear expression of her mounting rage. She began to pace in front of the wreckage.

Emotion: mad

Cluster: Anger / Hostility
PC1 (Valence): -1.94 Negative
PC2 (Disposition): 1.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.448
骂0.344
est0.340
🤬0.336
aggravated0.332
Suppressed:
a-0.495
de-0.485
la-0.387
H-0.344
latter-0.290