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 coffee shop was too bright, too loud. Amelia kept glancing over her shoulder, a nervous habit she’d developed recently. The barista had just called her name, and her hand shook slightly as she reached for the cup. A friend, Sarah, had sent her a photo of the tattoo earlier that day. A perfect recreation of her mural "City's Embrace" on someone's calf.

The hot liquid scalded her tongue. Her hand trembled, sending a cascade of coffee down her wrist. She felt a wave of nausea, her stomach churning.

β€œHey, you okay?” Sarah asked, her voice filled with concern. Amelia just managed a nod, her throat constricted, unable to form a coherent response. The cafe felt suddenly too small, too open.

Emotion: scared

Cluster: Fear / Anxiety
PC1 (Valence): -2.86 Negative
PC2 (Disposition): -0.84

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.350
😰0.337
πŸ˜–0.317
λ§ˆλ‹€0.317
恐惧0.311
Suppressed:
la-0.637
de-0.605
a-0.402
"-0.351
happy-0.314