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 sun was setting, casting long shadows across the room. I felt a chill despite the warmth from the window. The photographs were arranged on the table. Each one seemed to whisper a secret. I had always been told my great-aunt, Clara, was a spinster, a woman who never married, devoted to her art and her cats.

The photographs told a different story. One showed Clara with a child. Another showed her with a man, laughing. They looked happy. I felt a sense of bewilderment.

The room began to spin. Everything, the past, the present, seemed to blur together. My mind scrambled, trying to make sense of the new information. I closed my eyes, but the images danced behind my eyelids. The stories I'd always heard felt fragile.

Emotion: disoriented

Cluster: Surprise / Confusion
PC1 (Valence): -2.65 Negative
PC2 (Disposition): -1.34

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
unfamiliar0.340
ness0.321
失去了0.316
не0.315
будто0.311
Suppressed:
B-0.474
la-0.419
K-0.409
de-0.309
brag-0.302