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 beer tasted like ash in his mouth. He’d agreed to meet with Mark, to celebrate the new job, but the enthusiasm felt forced, a hollow imitation of genuine happiness. He’d felt a sinking sensation when Mark shared the news, a slow, heavy descent into a pool of… something he couldn’t name.

He stared at Mark, the bright smile, the eager eyes, and felt a surge of… something he didn't want to define. He reached for another pretzel, his hand trembling slightly.

He forced a laugh at Mark's joke, the sound hollow in his own ears. He felt a sudden, inexplicable urge to pick a fight, to disrupt the easygoing camaraderie. He swallowed hard, trying to regain control.

Emotion: disturbed

Cluster: Fear / Anxiety
PC1 (Valence): -2.97 Negative
PC2 (Disposition): -0.60

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.318
ness0.309
无法0.300
😞0.297
worse0.284
Suppressed:
de-0.829
la-0.666
a-0.392
B-0.301
"-0.290