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).

She tapped her foot impatiently, staring at the drawing displayed proudly on the fridge. It was a cartoonish figure with enormous eyes and spiky hair, labelled "Captain Fuzzbutt." It was her childhood friend, *exactly*. She felt a prickle of annoyance under her skin.

“Isn’t it cute, Auntie?” Lily chirped. The question was a challenge, she felt, a deliberate jab at her carefully constructed defenses. She kept her gaze fixed on the drawing. This was getting ridiculous.

She forced a chuckle. “Yes, dear. Very… imaginative.” The words tasted like ash. She wanted to rip the drawing down, to erase the evidence. She wanted to be left alone. She excused herself, claiming a headache, and retreated to her room, where she closed the door with a decisive click, finally shutting out the offending drawing.

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