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).

He was standing near the bar, chatting animatedly with the groom’s cousin. My eyes widened. I quickly ducked behind a very tall, very potted fern, hoping he hadn't noticed me. A prickle of heat crawled up my neck. I could feel my face flush.

"Everything okay, Sarah?" My best friend, Emily, her voice was laced with concern.

“Yeah, fine, fine,” I stammered, adjusting my dress and trying to appear casual. “Just… a bit warm, you know?” I fanned myself with my hand, attempting a smile that probably looked more like a grimace.

He was moving, heading towards the dance floor. The fern was no longer a sufficient shield. I considered feigning a sudden illness and making a hasty exit. My legs felt like lead.

Emotion: alarmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.304
😨0.272
worse0.269
😣0.261
не0.254
Suppressed:
de-0.323
enthusi-0.258
happy-0.235
delight-0.225
ية-0.217