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 evening air was thick with the scent of jasmine. Across the balcony, Leo was practicing his Vietnamese. She saw him from her vantage point, a half-empty glass of wine in her hand. The glow from his laptop illuminated his face, casting deep shadows on his features as he laboriously sounded out the words.

He looked up, meeting her gaze. He smiled sheepishly, waving the laptop in a mock apology. She shook her head, smiling back. The breeze rustled the leaves of the bougainvillea, and the sounds of the city drifted up to them.

She took a long sip of her wine, watching him. This was not the first, nor would it be the last time, she supposed, that she would watch him. She knew that the progress would be slow, and the path torturous, but she didn't mind.

Emotion: patient

Cluster: Passivity
PC1 (Valence): 2.71 Positive
PC2 (Disposition): 0.40

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
rest0.467
eventually0.415
leisurely0.369
慢慢0.357
a0.355
Suppressed:
无法-0.432
even-0.428
인해-0.421
無法-0.415
😣-0.391