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).

It was the sight of the manicured lawn that really got to Alex. The yard, which had always been a battleground of weeds and neglect, now looked pristine and perfect, as though it was never the site of so much anger and frustration. The sudden, unwelcome image ignited a fire within him.

He reread the listing, the glowing description of the "charming features" and "modern amenities" a stark contrast to the dilapidated, forgotten reality of his childhood. He had made a point of leaving it all behind.

He paced the confines of his small apartment, the comfortable space suddenly feeling suffocating. He went to the fridge and poured himself a whiskey, neat. The alcohol burned going down, but it didn't numb the hollowness in his chest. He felt…used.

Emotion: resentful

Cluster: Shame / Guilt
PC1 (Valence): -1.47 Negative
PC2 (Disposition): 1.50

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.485
aggravated0.263
愤0.261
👎0.245
L0.241
Suppressed:
own-0.318
soon-0.225
随后-0.225
serendip-0.219
🤩-0.210