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).

He stared at his reflection in the darkened computer screen, a stranger staring back at him. The lines etched around his eyes seemed deeper, the color drained from his face. The knot in his stomach was a physical manifestation of the knowledge. The information had been confirmed. The new hire, straight out of school, was earning more than he, after a decade of loyalty.

The screen reflected the harsh overhead lights, a sterile glow that mirrored the emptiness he felt. He ran a hand over his face, feeling the stubble of his beard. He hadn't shaved in days. What was the point?

He opened a drawer and took out his coat and keys, then walked out of his office. He didn’t say anything to anyone. He just left.

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