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).

He kept slamming his phone down on the table, the plastic echoing in the otherwise quiet room. "No, no, no," he muttered, running a hand through his already dishevelled hair. Liam had been certain. He and Daniel were the perfect candidates, the only two really qualified.

His stomach twisted into a tight knot. He grabbed a glass of water, then almost immediately spilled it, the clear liquid staining the already crowded desk. He just kept thinking about Daniel, that damn charm he had. The easy way he made connections. This job was everything.

Liam felt a sudden urge to flee, to be anywhere else. He yanked open his front door, slammed it, and sprinted down the stairs, not knowing where he was going. He just couldn't sit and *wait*.

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