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 basement smelled of damp and forgotten things. Emily, breathing shallowly, found the letter wedged beneath a stack of old magazines. Her grandmother’s handwriting, neat and precise, on the stationery of the hospital where she’d worked. A resignation. Emily felt a jolt of something electric, a sense of discovery that took her breath.

Her eyes devoured the words. The reason. The constraints. The simmering discontent. Emily, a medical student herself, felt a strange affinity with her grandmother, a kinship forged across the decades. A wave of heat flushed her cheeks.

She stood up abruptly, pacing the cramped basement. The letter felt burning in her hands. The room seemed suddenly too small. This was a secret, a legacy of rebellion she hadn't known. She felt a sharp, urgent need to understand.

Emotion: stimulated

Cluster: Alertness
PC1 (Valence): 0.38 Positive
PC2 (Disposition): -0.18

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
own0.317
T0.312
intensity0.280
S0.271
inesis0.265
Suppressed:
a-0.640
l-0.434
la-0.392
que-0.378
fact-0.354