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).

Eliza stared at the clutter-filled room, the walls seeming to close in on her. She felt depleted, like she was running on fumes. She was tasked with cleaning out her aunt's apartment, a difficult thing to do since her passing. The mere thought of it was exhausting.

She unearthed a series of letters and photographs, tucked away in an old trunk. One photo showed her aunt, a woman who always wore a stern expression, laughing while riding a rollercoaster with a man whose face she'd never seen before. Her aunt had spent her life dedicated to her family and its traditions. She had no time for frivolity. This image shattered the rigid construct she had built around her. Her hands felt clammy.

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