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 clock on the microwave in the breakroom seemed to taunt her. Each passing second ticked by at a snail's pace. “And then what do I do with the, uh… the thingamajigger?” It was the fifth time today he’d asked about the thingamajigger. Amelia rubbed her temples, feeling the beginnings of a tension headache. This process was a trial.

Her replacement, Kevin, was a well-meaning but utterly clueless individual. Every instruction was met with a blank stare, every task required multiple explanations. She felt a knot tightening in her chest. She had a mountain of her own work to do, and instead, she was trapped in this purgatory.

She gestured towards the filing cabinet with a stiff hand. "Put the thingamajigger... in the filing cabinet." Her voice was flat. She needed a vacation. A very long one.

Emotion: exasperated

Cluster: Anger / Hostility
PC1 (Valence): -0.99 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.435
l0.378
P0.331
que0.320
la0.269
Suppressed:
own-0.433
gradually-0.254
soon-0.248
unfolding-0.216
joyful-0.216