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 parking lot was already full, and the heat radiating off the asphalt felt oppressive. Michael had to get out of the house. The arguments with his wife had become a daily occurrence, each one escalating into a shouting match. He found himself pacing, unable to relax. He pushed through a crowd, his jaw clenched, scanning the chaotic scene.

His gaze snagged on a familiar shape at the bottom of a cardboard box. A bear, his childhood companion, Bartholomew. A jolt, a physical shock, ran up his spine. Bartholomew’s one ear drooped. He remembered his grandmother stitching it back together.

He moved toward the seller, a woman with kind eyes. “How much for the bear?” His voice sounded strained, even to his own ears.

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
느껴0.183
😣0.183
😖0.181
Suppressed:
a-0.260
de-0.258
la-0.256
happy-0.218
B-0.198