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 house was bigger than she thought. It had an old feeling, and a strange musty smell. Jessica moved in slowly. She wanted to savor the process of settling. She organized the kitchen, and then started working on the garden.

One afternoon, while trimming the rose bushes, she noticed a peculiar pattern of bricks in the wall of the house. She brushed away the dirt. Something did not make sense. She noted the anomaly.

Weeks went by, and the garden looked amazing. Then, she remembered the brick pattern. She tapped it. A section of the wall swung inward, revealing a small, hidden room filled with gardening tools and supplies. A perfect space. She sat down, and organized it.

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