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 kindergarten classroom was overflowing with brightly colored art and sticky-fingered children. She stood in the doorway, suddenly feeling the weight of the world on her shoulders. Mr. Henderson. His name, a low, ominous rumble in her memory. He had been a tough teacher, critical and quick to reprimand.

She watched as her daughter, Lily, ran towards him, her small face alight with excitement. She felt a pang, a sharp ache in her chest. She remembered Mr. Henderson's comments about her messy handwriting, his impatience with her creative ideas.

She took a step back, suddenly feeling out of place, a stranger in her own child's life. He would notice, wouldn't he? He would see her flaws reflected in her daughter.

Emotion: guilty

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.492
L0.398
탓0.354
been0.323
ashamed0.319
Suppressed:
de-0.400
(!)-0.279
la-0.276
😍-0.274
"!-0.264