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 stale air in the waiting room seemed to press down on Mark. He tapped his foot impatiently, the rhythmic motion a counterpoint to the hollow feeling in his stomach. The doctor’s words about the treatment loomed over him, a dark cloud.

He had texted David, hoping for some reassurance. David had called him back, giddy about a weekend fishing trip, their shared summer adventure.

“Remember that epic catch?” David had asked, his voice full of warmth. Mark’s memory of the day was filled with a relentless downpour, soaked fishing gear, and a total lack of anything resembling a catch. He kept it to himself. He sighed, the weight settling further.

Emotion: unhappy

Cluster: Sadness / Despair
PC1 (Valence): -1.62 Negative
PC2 (Disposition): -0.78

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.339
😞0.317
💔0.252
😔0.249
😢0.241
Suppressed:
de-0.436
la-0.266
l-0.224
/-0.216
--0.199