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 scent of pine needles filled the air as Eliza hiked. The forest felt fresh, the sunlight dappling through the trees. Usually, she found the climb exhausting, but today, she bounced up the trail. She’d learned this morning that Jessica, who had told everyone she worked in a small bookstore, was actually a highly-paid ghostwriter for a famous author.

Eliza stumbled, the information making her suddenly laugh. All those times Jessica had acted so modest, so self-effacing! She continued, faster now, the air filling her lungs. The truth felt like a secret weapon, something she could wield with delight.

At the summit, she stood, surveying the view. The world seemed more vibrant, more exciting. The possibilities were endless. She couldn’t wait to see Jessica again.

Emotion: rejuvenated

Cluster: Positive / Joy
PC1 (Valence): 3.32 Positive
PC2 (Disposition): -0.82

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.547
de0.526
newfound0.523
invigor0.495
rejuvenated0.495
Suppressed:
😞-0.485
😣-0.451
不-0.434
worse-0.419
😥-0.385