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 afternoon was too quiet, the air thick with an oppressive humidity. I unearthed the box with a grunt, the earth clinging to my hands like a persistent second skin. I didn’t know what I expected. Hope? Regret? I'd forgotten even doing this. Just more wasted energy on a past I'd much rather forget.

I flicked the lid open. A collection of silly knick-knacks and a letter. I shoved the letter into my pocket, figuring I'd read it later. A childish smile. A drawing of the house I grew up in. I didn't want to look at it. I kicked the box back into the hole, shoving the dirt back on top of it. I'd deal with this later. Right now, the only thing I wanted was to be left alone.

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