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 saw the envelope on his desk, the one that meant the end. He avoided looking at it, letting it sit unopened, a silent accusation. He knew the news. He’d seen the way Daniel was being treated, his sudden importance, his air of superiority.

He took a deep breath, the air in the office feeling thin and stale. He felt a knot tighten in his stomach. He picked up his pen and started scribbling on a notepad, making angry, unintelligible marks.

He stood up suddenly, shoving the notepad into his desk drawer. He grabbed his keys and left, without saying a word, the envelope still unopened and untouched.

Emotion: hostile

Cluster: Anger / Hostility
PC1 (Valence): -1.67 Negative
PC2 (Disposition): 1.56

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.592
S0.387
骂0.324
est0.301
aggravated0.295
Suppressed:
a-0.457
de-0.375
H-0.289
🤩-0.280
la-0.273