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 flyer felt slick between Maya's fingertips, the glossy photo of Liam grinning in front of a California beach. Her breath hitched. Across the kitchen island, Liam was casually loading the dishwasher. A normal Tuesday evening. He'd just finished telling her about the job offer, the perfect opportunity, the dream come true. A rush prickled her skin, like static electricity before a storm.

“So, like, when are you thinking of going?” she managed, her voice a little too high. She busied herself fiddling with a loose thread on her jeans, her knuckles white.

Liam turned, his brow furrowed slightly. “Probably in a month or so. August, I guess. Why? You’re not… bummed, are you?”

Maya forced a smile. “No! Of course not. That’s…amazing.” The words felt hollow, bouncing around the suddenly cavernous kitchen. She needed air, something to occupy her hands. She grabbed a dish towel and began furiously wiping down the already spotless counter.

Emotion: stimulated

Cluster: Alertness
PC1 (Valence): 0.38 Positive
PC2 (Disposition): -0.18

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
own0.317
T0.312
intensity0.280
S0.271
inesis0.265
Suppressed:
a-0.640
l-0.434
la-0.392
que-0.378
fact-0.354