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).

Agnes slammed the real estate agent's business card onto the gleaming granite countertop. The agent, a smirking man named Bradley, had the gall to suggest she "embrace the history" of her new home. *History?* It was a cramped, overpriced shoebox, and now apparently, it had a dead guy in the living room. She pushed a stray strand of grey hair behind her ear, her jaw tight. "Embrace the history," she muttered under her breath, a bitter taste rising in her throat.

She marched into the living room, the scene of the alleged "incident." A single, dusty sunbeam cut across the patterned carpet, highlighting the imperfections. She kicked a rogue throw pillow across the room. *Embrace this, Bradley, you pompous windbag*, she thought, her lips curling into a thin line. The thought of all the other people that had lived here before now felt like a bad joke.

Emotion: scornful

Cluster: Anger / Hostility
PC1 (Valence): -0.76 Negative
PC2 (Disposition): 2.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.406
C0.391
S0.276
I0.268
愤0.243
Suppressed:
own-0.396
unexpectedly-0.288
Suddenly-0.260
soon-0.256
serendip-0.256