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).

He paced back and forth in front of the building, each step a measured shuffle. The crisp autumn air did little to calm the tightness in his shoulders. He kept glancing at his watch. He felt as though he were reliving a nightmare, but this time, he was the one being forced to face it.

The receptionist called his name. He flinched. He stood there for a moment, unable to move.

He imagined the stories heโ€™d heard. He felt a sudden, inexplicable anger. Heโ€™d tried so hard to put the past behind him. Now, here it was, staring him in the face.

Emotion: worried

Cluster: Fear / Anxiety
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.285
S0.279
๐Ÿ˜ฐ0.277
๐Ÿ˜ฃ0.269
Worse0.250
Suppressed:
la-0.476
de-0.432
๐ŸŽ‰-0.233
happy-0.231
"-0.228