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’d been avoiding the internet for days, but the gnawing curiosity eventually got the better of him. The listing was a blow to the gut. The familiar porch, the crooked mailbox, even the specific shade of paint on the front door – it all triggered a wave of emotion.

He closed the laptop, the screen going black, but the image remained imprinted on his mind's eye. He felt a desperate urge to break something, to shatter the suffocating silence of his apartment. The air was thick, like a physical barrier.

He started to clean, a frantic, desperate attempt to regain some semblance of control. The small space of the apartment felt tighter, as though shrinking. The idea of going outside was exhausting, another obstacle.

Emotion: trapped

Cluster: Sadness / Despair
PC1 (Valence): -3.02 Negative
PC2 (Disposition): -0.80

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.573
ness0.401
无法0.366
hopeless0.360
😞0.353
Suppressed:
la-0.740
de-0.376
l-0.339
several-0.323
excitedly-0.322