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 new hire, Kevin, was a mirror image of him, or so it seemed. Same age, same background, same haircut. It was unnerving. Every shared joke, every similar mannerism felt like a ghost limb, a phantom ache. He found himself picking at his nails, the skin raw and bleeding.

His stomach churned as he showed Kevin the intricacies of the company's proprietary software. He wanted to tell him it was all a mess, a house of cards ready to fall. But he couldn't. He couldn't risk the accusation of insubordination. He felt trapped.

He kept having this urge to run, to just disappear, to leave everything behind.

Emotion: disturbed

Cluster: Fear / Anxiety
PC1 (Valence): -2.97 Negative
PC2 (Disposition): -0.60

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.318
ness0.309
无法0.300
😞0.297
worse0.284
Suppressed:
de-0.829
la-0.666
a-0.392
B-0.301
"-0.290