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).

Michael sprawled on the couch, the remote resting on his chest. The television murmured softly, a background to the quiet evening. Sarah was upstairs, humming as she prepared her bath. He smiled, thinking of her, her long hair, her easy laughter. Contentment settled over him like a warm blanket.

A chime echoed from the bedroom. It was her phone. He knew the ringtone. He recognized the number. The caller ID was a name he didn't know. Curiosity, a rare visitor, flickered through him.

He padded upstairs, feigning a casual air. "Everything good up here, love?" he called out, his voice slightly louder than intended. She emerged from the bathroom, a towel wrapped around her. "Just fine!" she responded with a smile. The phone, still chirping, was quickly silenced.

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