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).

"We're so sorry for the delay," the voice over the intercom intoned, a phrase that felt like a direct insult to Sarah. Her sister's graduation ceremony, a moment she'd been eagerly anticipating for years, was happening, or would have happened, without her. She chewed her lip, the taste of her own anxiety bitter. *Of course this would happen to you.* *You can't even get to the airport on time.*

She scrolled through social media, seeing the congratulatory posts, the photos of her sister beaming. A pang of envy shot through her, quickly followed by a rush of embarrassment. She pulled her sweater tighter around her, suddenly feeling cold. The bustling airport, usually a source of excitement, now felt like a prison.

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