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 stale air of the storage room clung to Ms. Ramirez. Boxes of old school supplies towered around her, each one a reminder of the life she was slowly losing. The recent divorce had left her adrift, and the relentless pressure of her job felt like a tightening noose. She felt a growing sense of detachment, a feeling that she was watching her own life from behind a sheet of glass.

A sudden light flickered on. A young woman, face beaming, stood there holding a canvas. "Ms. Ramirez! Remember me? It's Maria from your art class. I made this for you."

The painting was of a vibrant, sun-drenched landscape, bursting with color. It was almost too much, a stark contradiction to the suffocating grayness that had been enveloping her. The sheer exuberance of the artwork was almost painful.

Emotion: troubled

Cluster: Fear / Anxiety
PC1 (Valence): -2.20 Negative
PC2 (Disposition): -0.81

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.343
S0.317
ness0.302
😔0.294
😢0.272
Suppressed:
de-0.553
la-0.511
l-0.307
/-0.260
(!)-0.236