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 vibrant, buzzing atmosphere of the concert hall became a muffled roar. Clara’s ears rang. The music, a symphony she’d been anticipating all week, now felt alien and distant. Across the aisle, Daniel, the "renowned concert pianist," was admitting he was actually a piano tuner. A shiver ran down her spine, leaving her teeth chattering. She wanted to shout, to confront him, but her voice seemed to have abandoned her.

She pressed her palms against her cheeks, as if that would somehow contain the storm raging inside. Her breath came in short, shallow gasps. She felt a burning in her throat, a prelude to a sob she desperately fought to contain. She squeezed her eyes shut, trying to block out the world, the music, everything, but the revelation echoed in her mind.

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