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 champagne flute felt cool against Eleanor's palm as she surveyed the dance floor. Music thrummed, a happy pulse that vibrated through the floorboards and up into her feet. She tapped a toe, smiling as she watched her friends twirl in a joyous blur. The dress she’d chosen swirled around her legs in a way she found pleasing. It had been a good decision.

She spotted him across the room, leaning against a pillar and talking with her cousin. A flicker of... something… crossed her face, quickly suppressed. She took a slow sip of champagne, the bubbles tickling her nose. No need to let him disrupt the pleasant atmosphere.

"Having a good time?" a voice asked. It was Mark, a groomsman. She turned, meeting his eyes. "Absolutely," she replied, her voice steady and light. "The flowers are beautiful, and the cake looks divine."

Emotion: content

Cluster: Calm / Serenity
PC1 (Valence): 2.98 Positive
PC2 (Disposition): -2.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.446
affectionately0.435
😊0.392
😊0.386
thoughtful0.386
Suppressed:
S-0.682
L-0.537
C-0.485
😣-0.362
est-0.350