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 late afternoon sun cast long shadows across Mrs. Gable’s meticulously manicured lawn, the greenest in the neighborhood. Arthur, watering his prize-winning petunias, nearly jumped when she approached, her usually cheery face etched with an unusual seriousness.

"Arthur, dear," she began, her voice unusually low, "I was hoping, perhaps, I could… borrow a little bit of your yard?" He blinked, the water hose momentarily forgotten in his hand.

She gestured towards the sunny expanse of his back lawn with a hesitant hand. "Just a small patch, really. For a… a vegetable garden." Arthur swallowed hard, a strange pressure building in his chest. His petunias were demanding a drink.

Emotion: perplexed

Cluster: Surprise / Confusion
PC1 (Valence): -1.42 Negative
PC2 (Disposition): -0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤔0.330
L0.300
🤔0.296
S0.273
que0.264
Suppressed:
மகி-0.227
rest-0.222
ecstatic-0.212
alleviate-0.210
uplifting-0.196