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 sat ramrod straight in the armchair, listening to his daughter. "Dad, the family needs the money. The old violin is worth a fortune. Think of what you could do for your grandchildren!" His jaw tightened, the muscles in his face rigid.

He ran a hand over the smooth, worn wood of the instrument case resting on his lap. He could feel the familiar weight, the history. Each scratch and imperfection told a story. He felt a pressure behind his eyes, a familiar burning.

He stared out the window, watching the leaves swirl in the autumn wind. The violin was a piece of him, of his life, a reminder of his youth and his dreams. “It stays,” he said, his voice barely a whisper.

Emotion: stubborn

Cluster: Defiance / Spite
PC1 (Valence): 0.80 Positive
PC2 (Disposition): 3.73

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.780
l0.685
la0.605
either0.580
a0.501
Suppressed:
难以-0.435
终于-0.419
顿时-0.405
竟然-0.396
wondering-0.391