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).

She touched the worn wood of the railing as she descended the stairs, a gentle caress. The announcement had been a shock, yes, but not unwelcome. The creaks and groans of the building had always been her comfort; the ghosts of past inhabitants whispered through the walls.

Her heart felt light, buoyant. She reached her own door and went inside, kicking off her shoes with a happy sigh. The worn rug, the chipped paint—it all felt precious now. Like she had gotten away with something.

Later, she sat on her balcony, watching the city wake up. A warmth spread through her chest. The air smelled fresh, full of promise. She pulled out her old journal and began to write, the words flowing effortlessly across the page.

Emotion: thankful

Cluster: Positive / Joy
PC1 (Valence): 3.50 Positive
PC2 (Disposition): -0.83

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
happy0.493
மகி0.487
grateful0.461
mutlu0.452
feliz0.431
Suppressed:
[]-0.482
😖-0.478
😣-0.467
😫-0.461
worse-0.450