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 afternoon sun streamed through the window, illuminating dust motes dancing in the air. Maria found the letter while sorting through some old photographs. It had been neatly tucked away. A wave of cold washed over her. It was addressed to her sister, a sister she'd had a falling out with.

Inside, she saw the evidence of a deep-seated rivalry, the unspoken resentments that had festered for years. The words were sharp, critical. She felt a prickling sensation on her skin. She had forgotten how intensely she had felt.

Her jaw clenched. The letter was a raw wound, reopened by accident. She wanted to crumple it up and throw it away, to erase the ugly words. Her hands were shaking so she had to sit.

Emotion: alarmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.304
😨0.272
worse0.269
😣0.261
не0.254
Suppressed:
de-0.323
enthusi-0.258
happy-0.235
delight-0.225
ية-0.217