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 bustling city streets seemed to blur around Sarah. The cacophony of car horns, chattering voices, and construction noise felt like a physical assault. She kept her head down, avoiding eye contact, desperate to reach the hospital. The past few weeks were a relentless cycle of exhaustion, illness, and doctor visits. Everything felt difficult.

She almost bumped into a man on the sidewalk, mumbling an apology before hurrying on. Inside the hospital, she sat in the waiting room, trying to force herself to take a deep breath. She could feel the walls closing in. The clock seemed to slow, each tick a hammer blow against her patience.

Then, a nurse called her name, and the man she had almost collided with on the street came forward, as well.

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