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).

"I'm joining the Church of Scientology," his brother, David, had announced, a smug grin plastered across his face. Amelia felt a cold wash over her, a sudden, unpleasant dampness that seemed to seep into her bones. Her breath hitched. The room, usually a comforting space filled with the aroma of her mother's baking, now felt suffocating.

She played with the hem of her sweater, pulling at the threads, the action a frantic attempt to distract herself. Her eyes darted around the room, landing on the framed photos on the mantelpiece, each a carefully curated memory of their family. She suddenly felt a pressing need to run her fingers over each one, to reassure herself that nothing was truly changing.

Her throat felt tight, and she could hardly manage a whisper when she spoke. "But... but why?" she finally managed. The question sounded small, almost pathetic.

Emotion: unnerved

Cluster: Fear / Anxiety
PC1 (Valence): -2.36 Negative
PC2 (Disposition): -0.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
๐Ÿ˜ฐ0.302
๐Ÿ˜ฃ0.282
ะฑัƒะดั‚ะพ0.265
๐Ÿ˜–0.265
ness0.257
Suppressed:
de-0.585
la-0.495
a-0.403
happy-0.289
ecstatic-0.284