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 waiting room seemed to tilt. Walls were shifting, the muted floral wallpaper swimming before her eyes. She gripped the worn, faux-leather armrest, knuckles white, forcing herself to breathe. This couldn't be happening. Not here. Not now. The receptionist's voice, chirping a cheerful, "Dr. Harding will see you now," sliced through her thoughts like a rusty blade.

She stood slowly, legs feeling like jelly. The door to the office loomed. A deep breath. A hesitant knock.

The figure behind the desk, framed in the soft afternoon light, was familiar, yet impossibly alien. Mark Harding, the bane of her existence in elementary school, now a doctor. A therapist. A wave of nausea churned in her stomach.

“Come in, Sarah,” he said, voice smooth and practiced. It was a complete betrayal of the rough, taunting drawl she remembered.

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