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 musty air of the basement pressed in on David. The boxes, stacked high, threatened to topple. His mother’s legacy was a mountain of paper and memories. He’d come to find a specific item. Now, a different sensation began to take hold.

He reached for a stack of old creative writing assignments. His eyes flickered over the handwriting, the early attempts at storytelling. He imagined the sale. He imagined the price. He felt a growing need to take as much as possible.

He pictured the money he could get. A new car. A vacation. All within his grasp, if he played his cards right. The thrill of anticipation sent a shiver down his spine. He began to organize, to plan, to anticipate the reward.

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