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 waiting room buzzed with the low thrum of nervous energy. Amelia bounced on the balls of her feet, her hands clasped behind her back, a silly grin plastered across her face. The floral wallpaper seemed to bloom anew with every blink. She'd been waiting for an hour, but the delay felt like a secret, a thrilling little pause before… something.

A man, slouched in the corner, coughed, a dry, hacking sound. Amelia, without thinking, mimicked him, exaggerating the rasp, letting her tongue peek out as she did. His head snapped up, a flicker of surprise in his eyes.

"That's quite a show," he chuckled, a low rumble that vibrated through her.

"Just preparing for my starring role," Amelia chirped, feeling a lightness in her limbs that was almost giddy. She winked. "You?"

Emotion: playful

Cluster: Positive / Joy
PC1 (Valence): 4.42 Positive
PC2 (Disposition): 1.44

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
B0.662
(!)0.566
!0.562
!?0.553
😜0.553
Suppressed:
ness-0.685
😞-0.590
S-0.562
😔-0.470
идти-0.461