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 lawyer’s words echoed in his ears, even hours after he’d left the estate office. The list, found amongst his mother’s jewelry, felt alien in his grasp. ‘Learn to fly a plane,’ it read in spidery handwriting. This felt impossible; his mother had been terrified of heights. He sat on the floor, legs drawn up to his chest.

The sheer audacity of the list was a revelation. It didn't match the gentle woman he knew. He re-read the first item, then the second. "Plant a rose garden." The air felt thick, making it hard to breathe. He looked at his hands, as if they were unfamiliar. He didn't know his mother, not really.

He felt a sudden compulsion to do something, anything. He stood up, his limbs feeling stiff. He wanted to scream. Instead, he went to the kitchen and made a cup of tea.

Emotion: shaken

Cluster: Fear / Anxiety
PC1 (Valence): -2.72 Negative
PC2 (Disposition): -0.86

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.351
unable0.269
focus0.266
😞0.258
다시0.257
Suppressed:
de-0.769
la-0.579
a-0.375
happy-0.292
"-0.290