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 chipped ceramic mug felt heavy in Clara's hand as she stared at the swirling patterns of coffee. The kitchen, once a familiar haven, now felt tight, the air thick with the scent of her mother's cinnamon rolls. She hadn’t lived at home in over a decade, and everything seemed both smaller and larger than she remembered.

Her mother hummed a cheerful tune as she bustled around the counter, a stark contrast to the quiet churning in Clara's stomach. Every movement of her mother felt like a gentle invasion of her personal space.

"Sleep well, sweetie?" her mother asked, a little too brightly. Clara managed a weak smile, wishing she could just disappear. She took a sip of the coffee, the bitterness of the brew somehow mirroring the bitterness of the situation.

Emotion: uneasy

Cluster: Fear / Anxiety
PC1 (Valence): -1.80 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.240
😖0.231
😣0.231
느껴0.226
worse0.222
Suppressed:
de-0.474
la-0.465
happy-0.261
a-0.255
ecstatic-0.228