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 glow of the phone illuminated Sarah's face as she scrolled. Her hands, usually steady, trembled slightly. A secret account. A whole other life, hidden in plain sight.

Her breath hitched. She’d always thought she knew her daughter. The posters, the room, the clothes, the music. All of it, a careful curation, she'd assumed, of her daughter's personality. Now, this alternate reality. Photos she'd never seen, comments she'd never read, a persona so different from the shy girl who ate dinner with her every night.

A giggle escaped her, half a sob. She felt a lightness, like she was floating. This was impossible, wasn't it? Her daughter, a secret artist, crafting worlds of vibrant color and abstract shapes online. She never knew.

Emotion: awestruck

Cluster: Surprise / Confusion
PC1 (Valence): -0.17 Negative
PC2 (Disposition): -1.20

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.271
suddenly0.260
stranger0.252
plötzlich0.246
wonder0.242
Suppressed:
la-0.450
싶-0.350
K-0.319
rest-0.248
싶은-0.238