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 aroma of stale coffee and hot dogs from the yard sale vendors was nauseating. He had to clench his jaw, to keep from screaming. The bear, perched atop a stack of dusty books, seemed to observe him with a detached, knowing gaze. He felt exposed, vulnerable. The seller, a woman with a kind but vacant expression, smiled. "Isn't he cute? Reminds me of my own childhood." The sentiment, delivered with such casual ease, grated on his nerves. He felt a sudden, irrational urge to destroy something, anything. He forced a smile, his muscles aching with the effort. He turned and walked away, his chest tight, his breathing shallow.

Emotion: tormented

Cluster: Fear / Anxiety
PC1 (Valence): -3.33 Negative
PC2 (Disposition): -0.70

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
无法0.406
😞0.399
每一次0.378
無法0.369
S0.353
Suppressed:
la-0.803
de-0.755
a-0.407
B-0.404
l-0.354