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 renovation project had been a nightmare. Sarah leaned against the crumbling brick wall, the dust coating her skin. Each muscle ached. Years of living and dying from someone else's pain was beginning to take its toll.

The contractor had found something during demolition, a small metal box hidden behind the fireplace. Inside, a faded newspaper clipping and a photograph. The newspaper headline screamed of a murder. The photograph showed a woman, smiling, in front of this very house.

She closed her eyes, the image seared into her eyelids. She remembered her first encounter with the house. She had hoped to make it a home. She felt drained, as if the very air had been sucked from her lungs. She just wanted to go to bed and sleep for a week.

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
à°­à°¿-0.214