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 world felt brighter, somehow. Olivia had been waiting for this. Mr. Davies, her art teacher, had finally followed her. The evidence was irrefutable. Her meticulously crafted Instagram feed, a gallery of stunning portraits and avant-garde still lifes, had clearly caught his attention.

She arrived at school early the next day, her sketchbook tucked under her arm. The art room was deserted. She found Mr. Davies working on a large canvas, a subtle smile playing on his lips. As he looked over her shoulder, she knew he saw the potential she felt for herself.

Olivia deliberately walked past his desk a few times during the lesson, her movements graceful. Each time, she made eye contact and offered a small, knowing nod. She sensed she was the star pupil now.

Emotion: smug

Cluster: Positive / Joy
PC1 (Valence): 3.43 Positive
PC2 (Disposition): 1.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de1.109
l0.615
la0.585
😎0.570
😉0.564
Suppressed:
S-0.579
own-0.486
💔-0.483
😞-0.464
😣-0.455