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 paced his cramped dorm room, the worn rug doing little to soften his steps. Each stride was forceful, a tight, rigid movement. His jaw was clenched so hard he could feel the muscles in his face begin to ache. The rejection notice sat crumpled in his hand, a physical manifestation of his discontent.

“Unbelievable,” he growled, the word a low rumble in his chest. His gaze landed on his meticulously organized desk. He swept everything off – textbooks, pens, the framed photo of his family. The items clattered as they hit the floor. He ignored the mess.

He kicked a discarded water bottle across the room, the plastic catching against the wall and making a harsh sound. He wanted to break something, to obliterate the source of his current misery. He wanted to scream.

Emotion: mad

Cluster: Anger / Hostility
PC1 (Valence): -1.94 Negative
PC2 (Disposition): 1.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.448
骂0.344
est0.340
🤬0.336
aggravated0.332
Suppressed:
a-0.495
de-0.485
la-0.387
H-0.344
latter-0.290