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 porch swing creaked rhythmically as Eleanor gently pushed off the flagstones with her bare feet. The humid air of the summer evening hung heavy, but a gentle breeze stirred the honeysuckle vines that climbed the trellis, bringing a sweet scent. She sipped her iced tea, the condensation on the glass cool against her palm. A small, contented sigh escaped her lips. The day had been long, but now, watching the fireflies begin their dance across the lawn, a sense of well-being settled over her. The drawings her niece, Lily, had sent were sweet, and a bit odd, though. There was this figure... a tall, shadowy man with a crooked 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