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 tightness in Daniel's chest had vanished. He felt… free. The house felt brighter, more open. He’d just finished a long and grueling therapy session, and had found himself at his childhood home.

His mother, a woman of meticulous habits, had kept a massive archive of his school work. Every book report, every math test, every crayon drawing was in perfect condition. It was overwhelming, a bit bewildering, but mostly...joyous.

He ran his hand over the worn covers of his old textbooks. He thought about his mother’s dedication, and he smiled. It felt like a weight he had been carrying for years had been dissolved. He felt like he could fly.

Emotion: refreshed

Cluster: Positive / Joy
PC1 (Valence): 3.31 Positive
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.998
de0.703
!).0.384
(!)0.381
!".0.379
Suppressed:
S-0.782
C-0.478
L-0.414
😣-0.388
😞-0.376