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 letter lay on the mahogany table, catching the late afternoon sun. Clara hesitated, hand poised over the embossed name. She'd carried this ache, this weight, for so long. Now, a strange lightness seemed to fill her, a tingling sensation in her limbs. She opened the letter.

She felt a weight she hadn't realized she was carrying, slip away. She closed her eyes for a moment, letting the sunlight warm her face. The air in her small apartment seemed to vibrate with a new energy. She walked over to the old piano, her fingers finding the familiar keys.

She played a soft melody, a tune she hadn't touched in years. The music flowed through her, a release, a cleansing. She felt a profound sense of… anticipation. She looked out at the street below, seeing the world with fresh eyes, newly aware.

Emotion: hopeful

Cluster: Positive / Joy
PC1 (Valence): 3.68 Positive
PC2 (Disposition): -0.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.715
la0.549
B0.468
🤩0.452
enthusi0.433
Suppressed:
S-0.603
不-0.401
😞-0.399
worse-0.394
C-0.386