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 library, once a sanctuary, now felt like a prison. Ethan, sprawled across a chair, found himself unable to focus on his research. The silence, normally so welcome, now seemed to press in on him. He flipped through the pages of the book without reading, a growing sense of frustration building within him. He kept looking out the window, seeing other students going about their day. He wanted to join them.

A message. It was a link to a website. His college essay was now featured as an example. The irony was almost too much to bear. The words, once full of ambition, now felt like a reminder of a path he was no longer sure he wanted to follow. He slammed the book closed.

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