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).

Her flight to the industry conference was postponed. Instead of disappointment, a giggle escaped her lips. The conference was a required activity that she was not looking forward to. The thought of it had been making her anxious. She imagined the long-winded speeches and awkward networking sessions. Now, she wouldn't have to.

She sat in the waiting area, a small hop in her chest. She fished out a small notebook and pen and began to scribble down ideas for the novel she'd been working on.

She looked up at the passengers and smiled, a sense of well-being enveloping her. It was a lovely turn of events.

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
๐Ÿคฉ0.495
Suppressed:
S-0.896
๐Ÿ˜ž-0.430
ness-0.410
ใใ‚Œใงใ‚‚-0.387
L-0.379