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).

"You're moving from center to guard, Ben," the football coach had said. Ben, usually so self-assured, felt a knot of anxiety tighten in his chest. His heart thudded in his ears.

He paced the sidelines, the pre-game energy of the stadium a jumble of noise. The other offensive linemen looked calm and confident. He felt a wave of nausea.

He was built for the center position, for the snap, for the precision of the play call. Now he needed to start again. He had to block a different side.

He wiped his sweaty palms on his uniform, his mouth dry. He took a deep breath, trying to steady himself. He needed to be strong. For his team.

Emotion: rattled

Cluster: Fear / Anxiety
PC1 (Valence): -2.29 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.264
😣0.261
worse0.256
😞0.248
😖0.247
Suppressed:
de-0.571
la-0.414
a-0.377
happy-0.259
enthusi-0.234