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).

Michael stretched, yawning widely, the comfortable silence of their apartment a soothing balm to his mind. The Sunday paper lay spread across the coffee table, and the smell of Sarah's cinnamon rolls drifted from the kitchen. He felt a deep sense of well-being, a quiet joy that settled in his chest like a warm weight. Sarah called him to breakfast, and as he sat down, Sarahโ€™s phone rang. She answered, moving away from the table and speaking in a low voice. Michael watched her back, a slight frown touching his brow, but then he quickly pushed it away. He picked up his fork and started to eat, the cinnamon rolls still warm and delicious.

Emotion: content

Cluster: Calm / Serenity
PC1 (Valence): 2.98 Positive
PC2 (Disposition): -2.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.446
affectionately0.435
๐Ÿ˜Š0.392
๐Ÿ˜Š0.386
thoughtful0.386
Suppressed:
S-0.682
L-0.537
C-0.485
๐Ÿ˜ฃ-0.362
est-0.350