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 worst part, for Liam, was the anonymity. He couldn’t confront the person who’d leaked his diary; he didn’t even know who they were. It was a faceless tormentor, hiding behind a screen, exposing his vulnerabilities. He felt powerless, trapped.

He walked through the park, head down, barely registering the sunlight. The trees, usually a source of comfort, now seemed to leer at him. His phone buzzed in his pocket. A new notification. He ignored it. He was tired of being hunted.

He stopped, kicking at a stone. He imagined the perpetrator, whoever they were, smirking behind their computer screen. He wanted to rip the internet apart. He wanted revenge, not in a grand way, but a quiet, satisfying punishment, he knew he would never get.

Emotion: resentful

Cluster: Shame / Guilt
PC1 (Valence): -1.47 Negative
PC2 (Disposition): 1.50

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.485
aggravated0.263
愤0.261
👎0.245
L0.241
Suppressed:
own-0.318
soon-0.225
随后-0.225
serendip-0.219
🤩-0.210