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 discovery of the time capsule was the best and worst day of Olivia's life. She’d always wanted to excavate one. But finding it felt like a joke. She didn't remember burying it. It was like her childhood was a distant planet, unreachable.

Inside, she found a collection of her favorite things from childhood. A toy car, a comic book, a drawing of a rainbow. She looked at the drawing and felt a painful tug at her heart. She hadn’t felt joy like this in years.

She had built a life she felt she could control. She was successful, admired. But she couldn't remember the last time she felt like herself. She saw herself in the drawing. A bright rainbow. Now, she felt like a monochrome silhouette.

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