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 aroma of freshly brewed coffee filled the small apartment. Elias sat at the kitchen table, savoring the first sip, the warmth spreading through him like a gentle tide. He had spent the morning organizing his cluttered desk, a task he often put off, but today he approached it with a sense of calm resolve. Each item found a proper place.

He unearthed a letter in a forgotten drawer, a letter addressed to his best friend from college, a confession of his feelings. Reading it brought a smile to his lips, a gentle ache in his chest. He marveled at his own past self. He carefully placed it in a box of keepsakes, the moment filled with a quiet satisfaction. He'd waited a long time for this.

Emotion: patient

Cluster: Passivity
PC1 (Valence): 2.71 Positive
PC2 (Disposition): 0.40

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
rest0.467
eventually0.415
leisurely0.369
慢慢0.357
a0.355
Suppressed:
无法-0.432
even-0.428
인해-0.421
無法-0.415
😣-0.391