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).

“This is it,” whispered Maya to herself, her voice barely audible. She sat hunched in a chair, her shoulders tight. The air in the waiting room seemed thick, suffocating. The interview for the data analyst position, a role she and her best friend, Sarah, had both dreamed about for years, was about to begin.

She closed her eyes, trying to visualize a positive outcome. But all she could conjure were images of the endless spreadsheets she'd been studying. She felt a profound weariness, an unwillingness to confront the challenges ahead.

The door clicked open, and the interviewer called her name. Maya felt her body stiffen, as if a great weight had descended. She knew she had to try, but the effort required felt immense.

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