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 afternoon sun cast long shadows, but the warmth did nothing to penetrate the chill that settled in Anna’s bones. She’d spent the day in a fugue state, the world around her appearing flat and colorless. Now, she found herself in her boyfriend, Ben's, study, drawn to a hidden corner. She found his notebook on the desk, the cover inscribed, “The Ghost in the Machine.”

Her eyes scanned the pages, her stomach twisting. It was about her. The story described her anxieties, her vulnerabilities, her obsessive need for control. He'd even included her secret childhood nickname, a detail she'd guarded fiercely. She felt as though a veil had been ripped away, leaving her exposed and vulnerable. A sense of profound loneliness washed over her.

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