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 slammed the locker door, the metal echoing in the otherwise quiet room. The smell of sweat and liniment, usually a comfort, now felt cloying, a suffocating blanket. “Wingback?” He spat the word out, disgusted. He was a striker. His job was to score. His purpose.

His shoulders tensed, each muscle screaming a protest he couldn't vocalize. This wasn't happening. This couldn’t be happening. He balled his fists, the blood rushing into his hands, making his knuckles ache.

He kicked a discarded water bottle across the floor. He pictured the coach’s smug face. He hated that smirk. He hated the way everyone else was so nonchalant.

He ran a hand through his hair, tugging at the strands. The jersey felt scratchy against his skin. He didn't want this. He didn't want to change.

Emotion: irritated

Cluster: Anger / Hostility
PC1 (Valence): -0.86 Negative
PC2 (Disposition): 1.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.370
S0.282
P0.264
😠0.248
aggravated0.233
Suppressed:
own-0.300
joyful-0.211
joyous-0.207
खोले-0.206
newfound-0.205