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 stared out the window, watching the leaves of the oak tree sway in the breeze. A constant, restless pull tugged at him, a feeling that kept him from settling. He'd been snapping at his wife, snapping at the dog, snapping at everything. A tap on the glass startled him.

It was Mr. Garcia, his neighbour. He held out his hand, palm up, revealing a small, dry seed. “Ben, my friend, could I use a portion of your yard for a garden?” He felt the blood begin to rush in his ears. The suggestion, the image of anything growing on his land, made him want to scream.

He shook his head, a curt, definitive motion. “No. I have other plans for the yard.” He turned away, needing to be anywhere but there. The thought of planting a single seed, of anything being changed, felt suffocating.

Emotion: restless

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.51 Negative
PC2 (Disposition): -0.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
own0.485
’0.363
S0.333
tìm0.316
destination0.272
Suppressed:
B-0.421
заслу-0.297
টুকু-0.211
arada-0.204
background-0.202