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 spilled coffee had soaked into his shirt, a visual representation of the mess he’d made of the client pitch. He felt a deep, profound sense of inadequacy wash over him, a familiar wave. David ran a hand through his hair, the gesture revealing his frustration.

Across the room, Emily, the intern, stopped by his desk, looking curious. "Hey," she said quietly. "Are you alright?"

David just shook his head, avoiding her gaze. He felt the heat rising in his cheeks.

"I went to Camp Evergreen," Emily said softly. "It was... a disaster. Completely. I remember one day they made me eat an entire bowl of beans."

David choked back a laugh, a small crack in the dam of his despair. He understood. He had his own bean-related stories.

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
πŸ˜”0.364
ness0.324
Suppressed:
de-0.675
la-0.536
(!)-0.323
!-0.298
l-0.288