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 world felt muted. Ms. Rodriguez moved through her classroom like a ghost, her energy drained away. The students' chatter, usually a comforting backdrop, sounded like a dull roar. Everything seemed to require far more effort than it should.

A student from her first year of teaching, now a successful architect, arrived at her office, unannounced. He simply handed her a small, leather-bound journal. "For your thoughts," he said, and left as quickly as he'd come.

She opened the journal. The first page contained a single, beautifully drawn sketch of the classroom where she had first taught, a reminder of a period of her life that was less burdensome and more full of promise. She began to write.

Emotion: droopy

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.15 Negative
PC2 (Disposition): -1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.303
无法0.252
いつも0.238
😔0.230
:(0.228
Suppressed:
l-0.360
P-0.356
B-0.339
de-0.334
!-0.239