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 article, with its bold, accusatory headline, was pinned to the pass-through window. Chef Isabella stared at it, her face a mask of disbelief. "Undercooked," the critic had written. "Lacking in ambition." She felt the blood draining from her face, leaving her chilled to the bone. She ran a hand through her hair, tangling the strands further.

The kitchen, usually a place of controlled chaos, felt alien, hostile. The clatter of pots and pans, the sizzle of the grill, the chatter of the staff – all noises became grating, an unbearable symphony of failure. She grabbed a towel and scrubbed the counter with unnecessary force.

Isabella felt a stinging behind her eyes and a tightness in her throat. She closed her eyes for a moment, trying to regain control, but the image of the critic's words – the judgment – was burned into her mind. She felt the tears welling up, threatening to spill over.

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
それでも0.373
Suppressed:
la-0.672
a-0.618
de-0.541
😎-0.370
😎-0.359