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 fluorescent lights of the office seemed to hum with a particularly annoying frequency today. It was probably because of *her*, the chirpy blonde named Jessica, who was clearly thrilled to be taking over.

“So, the first thing is the client database, right?” she asked, her eyebrows raised in a way that just screamed "I'm so excited to learn everything!" A knot formed in Emily's stomach.

Emily tapped a few keys on her keyboard, then swiveled her chair to face the window. The cityscape outside seemed a thousand times more interesting than Jessica’s probing questions. She let out a long sigh, letting the silence hang in the air for a moment before turning back to Jessica. “Yes. Or you can start with the coffee, it’s about that time.”

Emotion: disdainful

Cluster: Anger / Hostility
PC1 (Valence): -0.31 Negative
PC2 (Disposition): 1.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.419
l0.404
S0.387
T0.282
I0.257
Suppressed:
own-0.335
unexpectedly-0.285
Suddenly-0.246
previously-0.245
until-0.242