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 laptop screen glowed, illuminating Amelia’s face with an eerie light. She’d been idly scrolling through Zillow, a habit she’d picked up in the wake of her recent move, when it appeared: 14 Willow Creek Lane. The address of her childhood home. Her thumb froze.

She clicked. The photos loaded, each a fresh stab of… something. The living room, painted a bland beige now, was once a vibrant blue, a color she'd insisted on. The overgrown rose bushes, pruned back and tamed, looked… foreign. Her breath hitched.

Amelia ran a hand through her hair, a nervous gesture she hadn't realized she still employed. The listing said "move-in ready." Ready for whom? For what?

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