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 relentless sun beat down on Liam, amplifying his sense of listlessness. He wandered the deserted beach, the rhythmic crash of the waves doing nothing to soothe his inner turmoil. Each step was an effort, his feet dragging through the warm sand. The low hum of a distant speaker carried a familiar melody. It was his song, "Drifting Away," playing from a beachside bar.

The singer's voice was smooth, polished, the lyrics delivered with a casual indifference. Liam had poured his heart into that song. But he didn't recognize it. It was now a soundtrack to someone else's vacation. The music felt like a distant echo of a life he no longer recognized, a ghost of a joy he could barely remember. He just kept walking, away from the bar and toward the horizon.

Emotion: gloomy

Cluster: Sadness / Despair
PC1 (Valence): -1.68 Negative
PC2 (Disposition): -1.37

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.368
无法0.319
😔0.300
いつも0.270
쓸0.255
Suppressed:
la-0.432
l-0.419
de-0.398
B-0.294
!-0.259