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).

Michael stared at the CEO's hastily arranged video call projected on the wall. The room, usually buzzing with the clatter of keyboards, was now unnervingly silent. He felt a profound sense of… something akin to being underwater. The voice droned on, outlining the specifics of the deal, but the words seemed to bounce off a mental shield.

He absentmindedly rubbed his left temple, a dull throb beginning to build. He noticed the sweat beading on his forehead, and felt a strange pull to reach for the nearest object to steady himself. His gaze darted around the room, taking in his coworkers' faces – a collection of confused and uncertain expressions.

The announcement finished, and the screen flicked off. Michael remained immobile, the silence amplifying the erratic thumping in his chest. A co-worker said his name, and for a moment he couldn't even remember who he was.

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