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 train rattled gently along the tracks, the rhythmic clack a soothing lullaby. Alex watched the passing landscape, the fields and forests blurring into a green and brown tapestry. The journey had been long, but now, the destination was near, and a sense of calm washed over him. He leaned his head against the window, the vibrations massaging his cheek.

Suddenly, a glint of metal caught his eye. Beside the tracks, partially covered by dirt, a small box. He craned his neck, trying to get a better look. It looked ancient. He hopped off at the next stop and returned, pulling the object free. Inside, he found a collection of old letters, a small toy car, and a few small, shiny coins. He didn't recognize any of it. He couldn't help but smile.

Emotion: relaxed

Cluster: Calm / Serenity
PC1 (Valence): 2.55 Positive
PC2 (Disposition): -2.48

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.638
disfrut0.418
прият0.386
楽し0.368
affectionately0.368
Suppressed:
L-0.598
S-0.558
C-0.435
because-0.428
ness-0.379