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 metallic tang of fear filled Marcus's mouth. He found the empty space in his driveway. His breath caught in his throat. He’d been working late, a long, grueling shift. Now this. He could feel his pulse hammering in his temples, a relentless, insistent beat.

He fumbled for his keys, then dropped them, the sound echoing in the sudden, sharp silence. His shoulders slumped, his muscles becoming leaden. Each detail of the scene – the empty pavement, the yellow sign, the indifferent sky – seemed to mock his vulnerability.

His gaze flicked from his house, to the empty spot where the car had been, and back again. A wave of nausea surged through him. He sank down onto the curb, his body betraying him with a sudden, complete exhaustion.

Emotion: tormented

Cluster: Fear / Anxiety
PC1 (Valence): -3.33 Negative
PC2 (Disposition): -0.70

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
无法0.406
😞0.399
每一次0.378
無法0.369
S0.353
Suppressed:
la-0.803
de-0.755
a-0.407
B-0.404
l-0.354