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).

He found the laptop open, the glow of the screen illuminating the late-night darkness. Leo, a man accustomed to flaws, felt his already fragile confidence shatter as he scrolled through his partner’s draft. It wasn’t a love letter; it was a character assassination.

The character, named after him, was indecisive, prone to anxiety, and, most damningly, boring. Leo ran a hand through his hair, the gesture more frantic than usual. He felt a hot flush crawl up his neck.

He looked around the room, at the carefully curated space they shared, a space he had always felt out of place in. Was this what he represented? A meticulously constructed facade, hiding a void of personality? He bit the inside of his cheek, a familiar habit when he was struggling.

Emotion: self-critical

Cluster: Shame / Guilt
PC1 (Valence): -1.66 Negative
PC2 (Disposition): 1.28

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-critical vector is projected through the unembedding matrix.

Promoted:
L0.748
S0.662
ness0.425
ly0.374
탓0.351
Suppressed:
la-0.492
soon-0.304
(!)-0.301
secured-0.284
cautiously-0.280