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 pounding bass still thrummed in Amelia’s chest even after she’d finally muted the music. Mrs. Gable's face, etched with lines of annoyance, had appeared at her door, a tiny whirlwind of disapproval. Amelia’s fingers twisted the frayed hem of her oversized sweater. She mumbled an apology, her voice barely a whisper, unable to meet the older woman’s gaze. The volume had been obscene, she knew it. Now, the quiet felt like a punishment.

She’d retreated into her apartment, the sudden silence amplifying the emptiness. She ran a hand across the counter, picking at a loose crumb. The overflowing ashtray on the coffee table seemed to glare back at her. A long, shuddering sigh escaped her lips as she made herself clean the place.

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