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 air in the room was thick with unspoken tension. Michael read the notice again, the stark words burning into his brain. He dropped it onto the cluttered desk, a sudden, desperate urge to smash something coursing through him. He clenched his fists, knuckles white.

His reflection stared back at him from the dusty mirror, a stranger with haunted eyes. He’d been warned. He’d been given chances. He'd ignored them all. The late nights, the reckless spending, the gambling – it had all caught up with him.

He pictured the smug face of his landlord, a man he'd deliberately avoided for weeks. The anger quickly dissolved into a deeper, more profound sense of emptiness. He’d let his family down. He’d let himself down. He just sat, staring at the ceiling.

Emotion: regretful

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 0.76

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.801
S0.774
H0.513
悔0.479
been0.473
Suppressed:
de-0.407
--0.320
随后-0.317
(!)-0.310
oldukça-0.301