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).

"More juice, Aunt Emily?" little Maya asked, her face smeared with grape. Emily found herself staring at another drawing, this one taped to the kitchen cabinet, featuring a spindly creature with elongated limbs. It looked oddly familiar.

The feeling in her chest was like a knot, pulling everything inwards. She remembered when she was little, she hadn't wanted to share anything. She felt it again. She reached for the juice carton and poured a generous amount for Maya, watching her intently.

"You like that friend, huh?" Emily asked, her voice a little too high-pitched. Maya nodded, her attention completely focused on the juice. Emily's gaze drifted back to the drawings. It was all a bit much. The house was already cluttered. The toys. The attention. *Her* special friend… and now this child.

Emotion: greedy

Cluster: Defiance / Spite
PC1 (Valence): 1.44 Positive
PC2 (Disposition): 2.64

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de1.148
a0.538
an0.510
or0.479
来实现0.464
Suppressed:
S-0.426
想起-0.356
久し-0.355
再次-0.354
причиной-0.342