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).

He stretched languidly, the morning sun painting the room in a golden hue. The gentle hum of the refrigerator was a familiar lullaby. Liam had just finished a particularly satisfying run, the air crisp and clear in his lungs. He felt a quiet contentment settle over him as he read a message from “MysticMage.”

MysticMage, who always seemed to know the right words, had sent a simple good morning. The chat log glowed with their shared interests, inside jokes, and the comforting cadence of easy conversation.

A few hours later, a notification. MysticMage. “Hey! I have a question about the bookstore on Elm Street, you know, where we always said we would go.” A smile curved Liam's mouth; he knew exactly what she meant.

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