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 attic air tasted of dust and forgotten things. Clara tugged at the string of the single bare bulb, the weak light doing little to dispel the gloom. A stack of photo albums sat precariously on a moth-eaten trunk. Her grandmother had always told stories of a prim and proper upbringing, a life of debutante balls and charitable works. But as Clara flipped through the brittle pages, her breath hitched. A photograph of her grandmother, her face obscured by shadows, a cigarette dangling from her lips, in the arms of a man who was most definitely *not* her grandfather. A strange pressure built behind her eyes.

She slammed the album shut. The sound echoed in the stillness. A wave of nausea churned in her stomach. Why had she even started this? She needed to get out, to breathe. The scent of lavender from the garden below beckoned.

Emotion: overwhelmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -0.94

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.385
😫0.332
😖0.319
😣0.319
😩0.314
Suppressed:
la-0.576
de-0.496
a-0.297
happy-0.290
l-0.264