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).

"Well, well, well," said David, his voice flat. He looked down at the time capsule, unearthed during renovations. It seemed to have come from a completely different life. The metallic gleam in the bright sun was like another insult.

He’d always felt a sort of…obligation. To his family, to his business, to the image he’d cultivated. Now, looking at the contents—a crayon drawing, a dried flower, a mix tape with a childish playlist—he experienced a peculiar sense of detachment. He didn't recognize the person who buried this. It was a person he could never be again.

The mixtape had a note on it: "Don't let them change you." He felt a tightness in his chest and looked up at the blue sky. The wind made the contents of the capsule rattle, and it sounded like a warning.

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