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 morning sunlight sliced through the blinds, illuminating the dust motes dancing in the air. Evelyn stared at the card. It was a cheap, store-bought affair, with a cartoon cat wearing a party hat. It was from... someone. No signature. Her birthday was tomorrow. The thought sent a cold wave across her skin.

Evelyn threw the card across the room, the flimsy paper fluttering to the floor. She hated surprises, and this felt like a particularly nasty one. She had spent the last few years carefully constructing a wall around herself, a barrier against the world and its relentless disappointments. This was a violation.

She went to the bedroom and took a bottle of pills from the drawer. She stared at the small white tablets, the words "take as needed" swimming before her eyes. Needed what? She wondered. She swallowed a handful, chased them down with a swig of water. They would help.

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