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 silence of the country road was suddenly broken by the sharp crack of his own curse. John’s face was a mask of disbelief and displeasure. He hadn't seen this place in years, but in his mind’s eye, the old birch was still standing, a sentinel guarding the entrance to his childhood.

Now, a gaping void. A jagged stump. He felt his pulse quicken. The veins in his neck stood out. He wanted to punch something, to destroy something. Instead, he took a deep, deliberate breath, forcing himself to calm down.

He walked around the area, hands shoved deep in his pockets, his gaze sweeping over the scene. The world seemed to tilt on its axis. He felt a strange combination of anger, sadness, and a deep-seated feeling of being utterly let down.

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