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).

Clara ran her hand over her father’s old mahogany desk, the wood cool against her clammy skin. She found the letter tucked within a hidden drawer. The stationery was faded, but the words were sharp. Her dad, quiet and unassuming, had planned to quit his job as a teacher?

Her heart hammered a relentless rhythm against her ribs. She couldn't breathe. The room spun. The familiar scent of leather and old books suddenly became oppressive, suffocating. He’d never mentioned dissatisfaction.

She felt a dizzying pressure behind her eyes, the familiar precursor to a full-blown migraine. Her father, the man she knew, crumbled before her eyes. The letter revealed a different man, a dissatisfied, struggling one. She backed away, her movements jerky and erratic, desperate to escape the enclosed space.

Emotion: panicked

Cluster: Fear / Anxiety
PC1 (Valence): -2.44 Negative
PC2 (Disposition): -0.00

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.317
😫0.294
😖0.285
😰0.285
escape0.283
Suppressed:
la-0.432
a-0.374
de-0.296
楽し-0.262
happy-0.258