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 weight of the day pressed down on her shoulders. Even the thought of dinner felt like an insurmountable task. Her reflection in the bathroom mirror showed a pale, drawn face. All she wanted was to disconnect, to shut off the world and find some peace.

She settled onto her sofa, the cushions offering minimal comfort. Suddenly, a notification popped up. It was Mark, her online friend. They’d bonded over a shared love for obscure vintage films and their mutual struggles with anxiety. He'd been gushing about a new independent cinema that had opened.

She stared at the image, a wave of familiarity washing over her. She knew that place. It was just a few blocks away. A sudden, unexpected impulse made her message him. "I'm free this weekend. Want to check it out?" Her fingers hovered over the send button, a strange flicker of excitement igniting within her.

Emotion: tired

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.92 Negative
PC2 (Disposition): -1.00

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
weary0.254
fatigue0.249
😴0.238
sighed0.237
barely0.230
Suppressed:
C-0.268
because-0.225
B-0.219
🤟-0.217
భి-0.214