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 musty scent of the attic air always made Amelia’s nose itch. Today, though, it smelled of her mother. She knelt, dusting off a chipped wooden box. Inside, nestled among moth-eaten scarves and forgotten photographs, was a tattered notebook. Her mother's handwriting, looping and familiar, filled the pages. The first entry: “See the Northern Lights.” Amelia closed her eyes, a warmth blossoming in her chest. She had the means now. She would start planning immediately.

The next day, she booked a flight. The image of the swirling green lights dancing across the sky, a spectacle her mother had yearned to witness, played in her mind. A small smile tugged at her lips as she made the arrangements, a feeling of lightness, a happy spring in her step. This was more than just a trip; it was a conversation, a continued bond.

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