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 eviction notice lay on the kitchen counter, its stark white blankness somehow more insulting than any accusation. Clara stared at it, her lips pressed into a thin, tight line. She grabbed a mug and poured herself a cup of coffee. The dark liquid sloshed over the rim as her hand trembled.

She hadn't missed a payment. Not a single one. She checked the bank statements again, just to be sure, as if some unseen entity was playing a cruel joke on her. It wasn’t a joke. Her gaze drifted to the chipped paint on the cupboards. How dare they!

Her shoulders stiffened. She walked over and stared at the door. Her apartment. Her life. Reduced to a cold, heartless piece of paper. She would not let them win.

Emotion: scornful

Cluster: Anger / Hostility
PC1 (Valence): -0.76 Negative
PC2 (Disposition): 2.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.406
C0.391
S0.276
I0.268
愤0.243
Suppressed:
own-0.396
unexpectedly-0.288
Suddenly-0.260
soon-0.256
serendip-0.256