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 cold wind whipped through Eleanor’s coat as she entered the pawn shop, matching the chill in her heart. She felt a sense of… unease… that had been growing steadily. The upcoming wedding, the pressure from her family – it was all too much. She sought refuge in the dimly lit interior, hoping for some respite.

She examined the jewelry cases, her focus erratic. Then, she saw it. A ring, a simple platinum band with a small, heart-shaped ruby. It looked almost identical to the ring her grandmother had lost during a hospital stay. A wave of dizziness washed over her.

She pointed, her hand trembling. “How much?” she asked. She knew she had to have it, to own a piece of that memory.

Emotion: unsettled

Cluster: Fear / Anxiety
PC1 (Valence): -2.33 Negative
PC2 (Disposition): -0.65

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.382
😞0.309
ness0.301
😣0.256
😔0.248
Suppressed:
de-0.721
la-0.597
a-0.365
"-0.299
l-0.274