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).

"More tea, dear?" Eleanor asked, her voice laced with an easy warmth. She poured a generous amount, the clinking of the china a pleasant sound in the sunlit conservatory. The sight of Mr. Finch's gardener trimming the hedges of the impossibly manicured Finch estate, a hedge that concealed a swimming pool and tennis courts, had been truly gratifying.

The knowledge had shifted something within her. It was as if a weight she hadn't realized she was carrying had simply lifted.

The floral wallpaper seemed brighter, the cushions beneath her more comfortable. A gentle warmth spread through her chest. She had always admired the Finches, their air of calm assurance. Now, she felt a similar internal stability. She smiled, a genuine smile, a feeling of wellbeing blossoming.

Emotion: satisfied

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.31

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.518
happy0.465
enjoy0.439
relaxed0.430
enjoys0.429
Suppressed:
S-0.515
C-0.511
😣-0.510
worse-0.477
😖-0.453