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 faint smell of old books permeated the library as Chloe sat at a table, poring over her aunt’s bucket list. She felt a comforting stillness settle over her. The sound of the pen scratching on the page, was music.

*Visit the Louvre Museum.* Chloe pictured her aunt, her eyes wide with wonder, admiring the masterpieces. A quiet sense of connection filled her.

Item two: *Learn to play the piano.* Chloe imagined her aunt's fingers dancing across the keys, creating beautiful melodies. A sense of completeness washed over her. She knew what she had to do.

Emotion: peaceful

Cluster: Calm / Serenity
PC1 (Valence): 2.40 Positive
PC2 (Disposition): -2.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
прият0.408
disfrut0.407
gently0.390
affectionately0.379
thoughtful0.376
Suppressed:
de-0.541
S-0.493
L-0.396
'-0.388
worse-0.367