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 room was heavy with the scent of lavender and old paper. Emily, sorting through her grandmother’s things, found it: a letter, carefully preserved, addressed to a man her grandmother had never married. The woman’s name was on the letter, a different one, a name she never used now.

She read the words, filled with unspoken longing and regret. Her grandmother had poured her heart into this letter, revealing a world of lost possibilities. The words were filled with a poignant innocence. Emily felt a tightness in her chest.

She closed her eyes, imagining her grandmother as a young woman, full of hopes and dreams. She felt the sudden surge of a profound compassion. It was more than sympathy; it was a weight of understanding.

Emotion: sorry

Cluster: Sadness / Despair
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.36

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.604
L0.549
😞0.382
😔0.364
ness0.324
Suppressed:
de-0.675
la-0.536
(!)-0.323
!-0.298
l-0.288