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 restaurant was too loud. The music, the chatter, the clinking of glasses – it all felt like a deafening roar. Across the table, Liam was talking animatedly about his upcoming adventure, his face flushed with excitement. I nodded, trying to appear engaged, but his words felt hollow, like they were bouncing off a wall inside me.

I excused myself to go to the bathroom. The air in the small space felt stale and oppressive. I splashed cold water on my face, but the coolness didn't soothe the churning in my stomach. Looking at my reflection, I saw a stranger, a ghost of myself.

The world seemed to have tilted, leaving me off-balance. I wanted to scream, to lash out, to stop time. I went back to the table, and the feeling was worse.

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