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 silence in her tiny apartment was deafening. The email, with its curt words of denial, glared at her from the laptop screen. She shut the computer with a snap, the sound echoing unnaturally loud in the small space.

Her hands trembled as she reached for a cigarette, even though she'd quit a month ago. She lit it, the familiar nicotine rush doing little to quell the turmoil within. The smoke filled the room, a hazy cloud mirroring her clouded thoughts.

She felt a hot flush crawl up her neck. Her landlord was already behind on the rent. She needed this scholarship. Her whole future depended on this. Now she had nothing.

Emotion: ashamed

Cluster: Shame / Guilt
PC1 (Valence): -1.97 Negative
PC2 (Disposition): 1.06

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.376
ashamed0.344
ness0.327
逃0.297
😞0.295
Suppressed:
de-0.539
que-0.300
😍-0.288
(!)-0.269
🤩-0.242