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).

He’d envisioned cozy nights by the fireplace, the scent of fresh-baked bread filling the kitchen. Now? He could hardly bring himself to walk through the front door. The detective had been remarkably casual, almost flippant, about the whole thing. A gangland hit, she'd said. Closed case.

Leo ran a hand through his hair, tugging at the strands. He kicked a stray toy across the living room floor, a toy his daughter had loved.

He felt a tightening in his chest. This wasn’t the "character" he’d been looking for. This was… a desecration. He grabbed his keys and slammed the door behind him. He needed air, needed to get away from these walls, from this whole cursed situation.

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