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 doorbell’s chime sliced through the silence of the apartment, a sound that felt entirely alien. She hadn't expected anyone. Her hands, clammy and trembling, fumbled with the locks, the bolt sticking stubbornly before finally giving way. A delivery person stood on the porch, holding a bouquet of vibrant lilies. No card. Her breath hitched.

She slammed the door shut, the lilies clutched against her chest like a shield. The scent, previously inviting, now assaulted her senses, a suffocating sweetness. Thoughts, fragmented and chaotic, ricocheted in her mind, each one a tiny shard of glass. Who? Why? The questions looped endlessly, a relentless carousel of panic. She began to pace, the lilies shedding petals onto the polished wood floor, each one a tiny accusation.

Emotion: hysterical

Cluster: Fear / Anxiety
PC1 (Valence): -3.19 Negative
PC2 (Disposition): -0.45

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.401
胀0.401
점점0.396
😫0.377
γγ‚Œγ§γ‚‚0.373
Suppressed:
la-0.672
a-0.618
de-0.541
😎-0.370
😎-0.359