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 watched the dust motes dance in the sliver of sunlight that pierced through the gap in the curtains. Days blended together, a seamless monotony of grey. Even the vibrant colors of the world seemed muted.

A ringing of the doorbell. It was Mrs. Petrov, her face as bright and cheerful as the flowers she always planted. She stood there, a smile plastered on her face, and he just wanted her to go away.

"Good afternoon, Mr. Johnson! I wanted to have a quick chat about the cherry tree. The blossoms are beautiful, of course, but the fruit makes a mess, and…"

He wanted to scream. Instead, he just stared at his feet, at the worn, faded pattern of the rug.

"It's on your side, so I thought I'd ask…" Mrs. Petrov continued.

He nodded, unable to speak. The tree. The fruit. The mess. It was all noise. He felt a wave of overwhelming indifference. He finally managed to croak out, "Do what you need to do."

Emotion: depressed

Cluster: Sadness / Despair
PC1 (Valence): -2.06 Negative
PC2 (Disposition): -1.08

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.417
每天0.353
无法0.330
own0.311
😔0.309
Suppressed:
de-0.492
B-0.393
P-0.352
l-0.310
!-0.307