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 afternoon sun streamed into the living room, but the warmth did nothing to penetrate the chill that had settled over me. I stood by the overflowing bookcase where my mother had stored my school work. I was overwhelmed.

I pulled out a thick composition book, its cover worn and faded. It contained my high school diary. The entries were a raw, unfiltered glimpse into my teenage angst. The thoughts I’d tried to bury. My cheeks felt hot.

Each sentence sent a jolt of discomfort through me. Memories, long dormant, resurfaced. I wanted to retreat, to hide. I felt a tightening in my chest.

I placed the book back on the shelf, my hands trembling. I turned and walked toward the kitchen. I needed a glass of water, something to calm my racing heart.

Emotion: tense

Cluster: Fear / Anxiety
PC1 (Valence): -1.32 Negative
PC2 (Disposition): -0.53

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.213
😰0.191
느껓0.183
😣0.183
šŸ˜–0.181
Suppressed:
a-0.260
de-0.258
la-0.256
happy-0.218
B-0.198