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 chipped mug in Amelia's hand trembled. Liam was across the small kitchen table, casually recalling the fireworks display. He remembered it as a warm summer evening, filled with laughter and shared hotdogs. Amelia felt a cold spike through her chest. It was nothing like that.

"The wind was howling, Liam!" she finally blurted, her voice sharp. "We nearly froze. And the hotdogs were burnt to a crisp." He just looked at her, his eyebrows raised in amusement.

She slammed the mug onto the table, the ceramic clinking loudly. The entire memory, the sheer injustice of his revisionism, sparked a fire in her. Had he even been there?

Emotion: outraged

Cluster: Anger / Hostility
PC1 (Valence): -1.96 Negative
PC2 (Disposition): 1.32

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.467
恨0.334
🤬0.327
aggravated0.323
骂0.312
Suppressed:
H-0.315
☺️-0.286
optimistic-0.278
adventurous-0.262
मुस्कुरा-0.260