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).

"Oh, for goodness sake," she mumbled, clicking the link in the email. It was from her sister, a simple subject line: "Remember?" She remembered. She remembered the house. The smell of her mother's baking, the sound of her father's laugh, the creak of the stairs. Now, this… a listing.

Her jaw dropped open. She hadn’t expected this. She felt disoriented, like she’d stepped onto a different planet. The layout of the rooms felt foreign; the garden, once filled with her mother’s roses, now a manicured lawn. Her head swam.

She pressed her palms to her temples, trying to ground herself. The words of the description swirled around her: “Spacious living areas… ideal for entertaining…” Entertainment meant nothing to her right then.

Emotion: dumbstruck

Cluster: Surprise / Confusion
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
будто0.326
не0.285
😶0.275
仿佛0.267
依旧0.258
Suppressed:
enthusi-0.278
la-0.264
K-0.260
wholeheartedly-0.255
쾌-0.254