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 rhythmic clacking of the typewriter filled the small apartment, a familiar comfort. He was lost in the flow of words, the story unfolding before him. Then, a sharp knock.

It was Mrs. Davis, her face a mixture of annoyance and fatigue. "The tapping! It's incessant! I need to sleep!" Her voice was strained, but her eyes betrayed a deeper emotion.

He stopped typing, the words fading in his mind. He looked at the manuscript. He thought of his late wife, who had always encouraged his writing, a constant support. He realized he was just going through the motions.

He looked at the empty space where she used to sit. He felt a profound sense of loneliness. He closed the door. The silence in the apartment felt heavier than ever.

Emotion: sentimental

Cluster: Nostalgia / Reflection
PC1 (Valence): 0.89 Positive
PC2 (Disposition): -1.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.677
fondly0.567
memories0.503
曾0.488
reminis0.466
Suppressed:
either-0.515
or-0.488
[]-0.473
--0.408
не-0.396