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 parking lot swam before Sarah's eyes. The sun glared down, reflecting harshly off the asphalt. She felt detached, as if watching herself from a distance as she fumbled with her phone, searching for the school's directory. Ms. Thompson. It couldn't be. Not *the* Ms. Thompson.

Her palms began to sweat. She leaned against the car, suddenly exhausted. She’d pictured her daughter’s school experience very differently. A fresh start. New faces. The comforting distance of the unknown. Her mind filled with images of dusty chalkboards and the sharp, critical gaze of a woman whose voice she still heard in her nightmares.

A wave of nausea hit. She pressed her hand against her stomach, trying to steady herself. This felt like a trick, a cruel joke. She needed air. Needed to sit. Before she walked in there and did something she would later regret.

Emotion: disoriented

Cluster: Surprise / Confusion
PC1 (Valence): -2.65 Negative
PC2 (Disposition): -1.34

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
unfamiliar0.340
ness0.321
失去了0.316
не0.315
будто0.311
Suppressed:
B-0.474
la-0.419
K-0.409
de-0.309
brag-0.302