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 air inside the pawn shop was still and quiet, a stark contrast to the bustling city outside. Owen ambled in, his hands in his pockets, the day sprawling before him like an open book. He had no plans, no obligations, just the freedom to roam.

He browsed the various displays, pausing to examine a vintage camera, a collection of antique coins, and a stack of faded photographs. And then, he saw it. A ring, a simple gold band with a single, emerald stone. It was an echo of a time he thought he'd lost forever. His grandmother's ring.

"How much?" he inquired of the shopkeeper, a man with a tired expression. The price was manageable. He paid, a small, knowing smile playing on his lips. As he slipped the ring onto his finger, a feeling of deep satisfaction filled him.

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