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 routine was a cage. Wake up, eat, work, eat, sleep. Repeat. Every day, the same grinding machinery. He found the list while clearing out his mother's desk. The script was unmistakable, her loopy handwriting.

"Swim with dolphins," "Go skydiving," "Write a song." A sharp pang resonated within him, a feeling of being both surprised and saddened. His mother, always so pragmatic, had concealed a desire for adventure.

He stared at the words, at the unfinished dreams. A wave of apathy washed over him, leaving him feeling cold and empty. He sat down and stared blankly at his computer screen. The tasks before him seemed meaningless.

Emotion: depressed

Cluster: Sadness / Despair
PC1 (Valence): -2.06 Negative
PC2 (Disposition): -1.08

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.417
每天0.353
无法0.330
own0.311
😔0.309
Suppressed:
de-0.492
B-0.393
P-0.352
l-0.310
!-0.307