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 man’s ears rang. The words swam before his eyes. He stumbled back, bumping into the mailbox, knocking it askew. His face flushed red, and he felt a searing heat crawl up his neck.

He felt a deep, profound sense of betrayal. The world, usually so clear, became a blur. The carefully constructed peace he'd worked to build was being torn apart.

He wanted to run, to escape the scene of the crime. His heart pounded in his chest, a desperate, frantic rhythm. He wanted to scream.

He opened his mouth, but only a strangled sound escaped. He took a staggering step, then another, his legs feeling like lead.

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