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).

"Another rejection," Maya muttered, tossing the email onto the cluttered desk. Her apartment, once a vibrant hub of creativity, had slowly devolved into a space of dusty canvases and half-finished projects. She felt a weight bearing down on her, a pressure that squeezed the air from her lungs.

A direct message appeared on her Instagram. A photo of a tattoo. A direct copy of her "Ephemeral Bloom" series, delicately inked on a shoulder blade. A feeling of… constraint washed over her.

She tapped on the profile. A young woman, smiling. "Obsessed! Thank you for the inspiration!" The words felt like a violation. Maya had poured her heart, her dreams, into that piece. It was a cry for freedom and the fleeting beauty of life. It was now someone else’s property.

She felt the sharp sting of tears, not from sadness, but from a frustration she couldn’t name. She wanted to scream, to run, to smash something. Instead, she just sat, still, the digital image burning in her mind.

Emotion: stuck

Cluster: Sadness / Despair
PC1 (Valence): -2.17 Negative
PC2 (Disposition): -0.72

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.487
缺乏0.401
L0.361
每天0.348
stagn0.348
Suppressed:
l-0.472
la-0.346
随后-0.322
undamaged-0.268
C-0.263