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 email notification popped up on her screen, and she felt a sudden dryness in her mouth. The subject line, "Party Update," promised excitement, but the words within delivered a different message altogether. Her breath hitched. She scrolled through the list of guests, all of whom had been so excited to come. Now, nothing. She felt a growing pressure inside her chest.

She stood abruptly, pushing the chair away from the desk with a sharp scrape. She needed air. The office, usually a familiar and comforting space, suddenly felt suffocating. She walked to the window, staring at the cityscape, its vibrant lights now appearing blurred through a sudden misting of her eyes. She wanted to scream.

Emotion: anxious

Cluster: Fear / Anxiety
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.315
😖0.302
worse0.297
😰0.285
own0.267
Suppressed:
la-0.441
a-0.382
de-0.365
"-0.256
happy-0.256