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 stale air in the staff room felt thick with unspoken anxieties. Sarah, a science teacher, was drowning in administrative duties, curriculum changes, and increasingly demanding parents. Each day felt like navigating a minefield, and the thought of another parent-teacher conference made her stomach churn.

A brightly colored box, delivered to her desk, brought a momentary distraction. Inside, she found a collection of handmade bath bombs, each infused with a different essential oil. A note read, β€œTo the best science teacher. You always made learning fun! – Maya.”

The scent of lavender and citrus filled the room. Maya, the quiet girl who always had a smile for everyone. Sarah ran her fingers over the smooth surface of a bath bomb, imagining herself soaking in a hot tub, away from all the pressures. It was a fleeting, but welcome, escape.

Emotion: unhappy

Cluster: Sadness / Despair
PC1 (Valence): -1.62 Negative
PC2 (Disposition): -0.78

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.339
😞0.317
πŸ’”0.252
πŸ˜”0.249
😒0.241
Suppressed:
de-0.436
la-0.266
l-0.224
/-0.216
--0.199