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 cursor on the laptop screen seemed to mock her. The familiar listing, with its pixelated photos and optimistic descriptions, stared back at her. Her childhood home. For sale. A wave of nausea washed over her, a bitter taste rising in her throat.

She hadn't been back in years. A lifetime, really, since the move. She scrolled through the pictures, each one a fresh stab. The chipped paint on the porch swing, the overgrown rose bushes, the peeling wallpaper in her old bedroom. All reminders of a past she couldn't reclaim.

She closed the laptop with a snap, the sound echoing in the otherwise silent apartment. The sudden movement sent a tremor through her. She felt like a glass dropped on the floor; shattered and beyond repair.

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