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 world seemed to sharpen into focus. It was as if a hazy film that coated everything had suddenly cleared. "I've been thinking about this for a long time," said Aunt Carol, her voice barely above a whisper. She explained her decision to become a Wiccan. Her sister, my mother, was visibly upset, her face ashen.

I, however, found myself humming a little tune under my breath. The constant, unspoken disapproval that had always existed within the family, the undercurrent of judgment regarding my own spiritual beliefs, seemed, at last, to diminish. My shoulders relaxed, and I could finally breathe freely. I felt the faintest trace of a grin playing on my lips.

Emotion: relieved

Cluster: Positive / Joy
PC1 (Valence): 2.22 Positive
PC2 (Disposition): -0.97

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.638
finally0.606
finalmente0.597
relieved0.561
enfin0.546
Suppressed:
l-0.443
:(-0.439
😣-0.429
S-0.388
😖-0.385