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 aroma of freshly baked bread filled the small apartment, and Maya inhaled deeply, a soft smile spreading across her face. She’d just pulled a perfect loaf from the oven. Picking up her phone, she meant to text her sister a picture. A quick snap, a message of pure, unadulterated joy. Instead, the image, followed by a simple "Feels good," pinged into the phone of her stern, recently-widowed neighbor, Mr. Henderson.

A moment of horror, quickly followed by a rush of heat to her cheeks, threatened to overwhelm her. She expected a curt reply, or worse, silence. Instead, a message arrived: "Looks delicious, dear. Smells like heaven out here." Maya's jaw dropped.

Later that evening, a knock came at the door. Mr. Henderson, holding a small plate of his own cookies, stood there. He even cracked a small, fragile smile. Maya's chest felt light as she looked at him and the plate of cookies.

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