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 rhythmic click of the keyboard, the gentle hum of the server, the soft glow of the monitor – for Anya, her small home office was her sanctuary. After months of tireless work, the “Chrono-Tracker” was finally complete. A sense of lightness filled her, like a balloon ready to float away.

She was scheduled to present her project at the tech conference, and was surprised when another presenter, a man named Ben, approached her. He had something to say.

Ben described the exact same project she was working on. His words, delivered with a similar enthusiastic energy, sparked a feeling of immense pleasure within her. He was beaming. A sense of warmth spread through her, and she found herself mirroring his smile.

Emotion: happy

Cluster: Positive / Joy
PC1 (Valence): 3.98 Positive
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.676
excitedly0.533
enthusi0.505
de0.498
🤩0.495
Suppressed:
S-0.896
😞-0.430
ness-0.410
それでも-0.387
L-0.379