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 faded, yellowed paper felt brittle under Eleanor’s fingertips. She’d found it tucked away in a shoebox at the back of the closet, the words scrawled in her partner, David’s, familiar hand. “Swim with dolphins,” it read, followed by a list of other adventures. Her breath hitched.

She sat on the edge of the bed, the list a stark contrast to her own life, which lately had been a loop of work and quiet evenings. A wave of inadequacy washed over her as she thought about her own aspirations. Had she ever even made a list?

Her gaze flickered over the neat, precise handwriting. It seemed so purposeful, so vibrant, and then drifted back to her reflection in the mirror, where she saw a woman who felt increasingly…stuck. She crumpled the list in her hand, the paper crackling, then smoothed it out, a fresh wave of guilt rising.

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