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 late afternoon sun cast long shadows across the living room as David settled onto the sofa. The apartment, bathed in a golden hue, seemed to breathe with a quiet contentment. He reached for the worn wooden box tucked away on the bookshelf, his hand lingering on its surface before lifting the lid. The list was nestled beneath a collection of dried flowers and faded photographs.

His eyes scanned the familiar words, each one a brushstroke painting a portrait of the man he loved. “Visit every continent.” He imagined their future, hand-in-hand, exploring the world together. The thought settled in his chest, a feeling of deep serenity. “Write a novel.” He pictured him, hunched over his laptop, lost in the creative process.

Emotion: safe

Cluster: Calm / Serenity
PC1 (Valence): 1.86 Positive
PC2 (Disposition): -2.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
soon0.402
affectionately0.357
disfrut0.329
необходи0.325
🥰0.317
Suppressed:
S-0.474
C-0.416
L-0.408
worse-0.379
even-0.370