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).

He flinched, as if the needle itself was piercing his own skin. He found himself staring at the man, the lines of his artwork – his abstract patterns of emotion – now permanent on a bicep.

His breath hitched. He had spent years trying to control these shapes, this form. Now, the man, a stranger in a coffee shop, carried his chaos. He took a hasty sip of his coffee, as if to calm his nerves.

He felt unmoored, the world tilting slightly. It was as though a part of himself had been ripped away and put on display. He paid his bill without meeting the barista's eye and rushed out, needing to be anywhere but there.

Emotion: vulnerable

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.69

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.327
ness0.319
😞0.299
😔0.277
😣0.258
Suppressed:
de-0.567
la-0.488
l-0.271
(!)-0.267
(!)-0.257