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 air in the office crackled with tension, but Sarah found herself strangely at peace. She closed her eyes for a moment, taking a deep breath. She could feel the rhythm of her own heartbeat. She knew she had worked hard, pushed herself, and learned a lot.

The competition was fierce, with both her and Mark vying for the same position. She knew her chances, and she knew Mark's strategy. She felt a profound contentment just knowing she had grown as a person.

She reached for her mug of tea, the warmth seeping into her hands, a feeling of comfort washing over her. The challenges had helped her discover new strengths, and that, in itself, was a reward.

Emotion: fulfilled

Cluster: Positive / Joy
PC1 (Valence): 3.22 Positive
PC2 (Disposition): -1.74

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.819
மகி0.415
happy0.397
grateful0.389
de0.385
Suppressed:
S-0.599
worse-0.465
C-0.460
either-0.436
😣-0.426