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).

Sarah stared at the blinking cursor, the white space of the laptop screen a stark emptiness mirroring the chaos in her mind. Her mother’s retirement party was tonight, and she was supposed to give a speech. She’d been putting it off, convinced that inspiration would strike. Now, the clock was ticking. A bead of sweat trickled down her temple, and she swiped at it with a shaky hand. The floral wallpaper of her childhood bedroom seemed to close in, the scent of lavender potpourri suddenly cloying and suffocating. She took a deep, ragged breath, trying to calm the frantic thumping in her chest.

Her fingers danced erratically over the keyboard, deleting the few stilted sentences she’d managed to write hours ago. A wave of nausea washed over her. She got up abruptly, pacing the small space, each step feeling heavy and uncertain. The party was just hours away.

Emotion: overwhelmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -0.94

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.385
😫0.332
😖0.319
😣0.319
😩0.314
Suppressed:
la-0.576
de-0.496
a-0.297
happy-0.290
l-0.264