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).

Liam stared at the chipped mug, the lukewarm tea growing cold in its depths. His mother’s handwriting. He traced the words, “Climb a mountain,” on the list, a list he'd found tucked inside her knitting basket. The air in the flat felt thick, heavy. He hadn't been to the gym in weeks. She'd always been so adventurous.

He shuffled through the pages. “Learn to speak Spanish.” “Write a novel.” He pictured her, the gentle smile, the way she'd hum while she worked. He'd never know if she had wanted to write a novel or if it was just a passing thought.

A tremor ran through him. He felt the sting of unshed tears. The silence in the house, the empty chair, was a physical thing. He had to do something. Anything, to break the hold of the quiet despair.

Emotion: melancholy

Cluster: Sadness / Despair
PC1 (Valence): -1.78 Negative
PC2 (Disposition): -1.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.445
😞0.376
😔0.337
いつも0.327
sadness0.312
Suppressed:
de-0.739
la-0.548
l-0.497
/-0.314
!-0.290