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).

“About damn time,” muttered Grace under her breath, staring at the unfamiliar handwriting. The letter felt heavy in her hand, filled with a weight that had been absent from her life for years. She sat down at her kitchen table, the chipped surface a familiar comfort, and let it go.

She read the letter, the words echoing in the silence of her apartment. Her knuckles turned white. It was all a complete fabrication. His version of what happened. Her heart, which had just started to heal, was breaking open once again.

Grace went to the fridge and grabbed a bottle of wine. She drank it straight from the bottle, the liquid burning her throat, then opened another, and another, until she finally passed out.

Emotion: hateful

Cluster: Anger / Hostility
PC1 (Valence): -1.71 Negative
PC2 (Disposition): 2.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.553
S0.367
retali0.313
aggravated0.310
恨0.308
Suppressed:
la-0.334
own-0.310
optimistic-0.270
been-0.264
gradual-0.262