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).

My sister, Maya, had always been the free spirit. But this… this was different. I paced the worn carpet of my childhood bedroom, the news still ringing in my ears. Buddhism. She was becoming a Buddhist.

The air felt thick, charged. I gripped the edge of the window frame, knuckles white. The late afternoon sun cast long shadows, painting the room in stripes of gold and darkness. I remembered Maya’s laughter, her silly jokes, her love for our grandmother’s Sunday afternoon sermons.

“But… Grandma…?” I had asked, my voice a strained whisper when she’d told us. Maya had merely shrugged, her face calm, almost serene. This didn't make sense.

Emotion: perplexed

Cluster: Surprise / Confusion
PC1 (Valence): -1.42 Negative
PC2 (Disposition): -0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤔0.330
L0.300
🤔0.296
S0.273
que0.264
Suppressed:
மகி-0.227
rest-0.222
ecstatic-0.212
alleviate-0.210
uplifting-0.196