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 envelope crumpled in her grasp as she reread the paragraph, the words twisting and turning into an incomprehensible jumble. The neatly typed sentences had morphed into something alien, a cruel joke. Her throat constricted. She struggled to swallow, as if a physical barrier had suddenly materialized there. She dropped the envelope. It landed on the pristine white tablecloth, like a fallen petal.

She wanted to scream, to lash out, but the words caught in her throat, strangled by the sheer impossibility of what she'd just read. She felt a tremor run through her body, starting at her toes and ending in her jaw, which began to tremble. Her eyes darted around the room, as if searching for a hidden camera, or a clue that it was all a prank.

Emotion: dumbstruck

Cluster: Surprise / Confusion
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
будто0.326
не0.285
😶0.275
仿佛0.267
依旧0.258
Suppressed:
enthusi-0.278
la-0.264
K-0.260
wholeheartedly-0.255
쾌-0.254