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 late afternoon sun slanted through the grimy window, illuminating dust motes dancing in the air. The notice lay face up on the cluttered desk, a pale square of judgment. The cheerful music from the neighbor's apartment seemed especially grating.

She felt a wave of nausea wash over her. She knew this was coming. Still, the reality of it felt like a punch. She slumped into the chair, the creaky springs protesting her weight. She had no one to turn to.

Her eyes welled, and she blinked rapidly, trying to keep the tears at bay. She thought of her prized collection of books, of the small, carefully curated world she had built within these walls.

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