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).

Dinner was awkward. My wife, Lisa, announced she was attending a weekly Wiccan circle. I choked on my water.

I stared at her, trying to read her expression. She looked… serene. That was the problem. I felt a rising tide of suspicion, an unfamiliar sensation.

Her explanation felt like a foreign language, filled with terms and concepts I didn't grasp. I picked at my food, the flavors suddenly bland.

As she spoke, I focused on the way the light hit her hair, the curve of her jaw. The world shifted. I excused myself. I needed some space. I found myself standing in the cold night air, unable to fully relax.

Emotion: skeptical

Cluster: Suspicion / Vigilance
PC1 (Valence): -0.02 Negative
PC2 (Disposition): 1.45

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.468
que0.396
la0.346
a0.333
🧐0.327
Suppressed:
再び-0.304
终于-0.299
再次-0.270
終於-0.263
momentarily-0.255