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).

"Just another rejection," she mumbled, stuffing the email back into her bag. Clara slumped onto a park bench, the vibrant colors of the autumn leaves seemed to mock her bleak mood. She felt like a faded photograph. Then, her hand brushed against something bulky on the ground. A wallet.

She picked it up. Inside, a significant sum of cash. More than enough to pay off the mounting bills, the overdue rent. A wave of conflicting emotions crashed over her. She imagined the relief, the freedom from the constant financial strain. She could finally breathe.

She felt a hollowness inside, a vast chasm that the money couldn't fill. She pictured herself, alone, adrift. And she wondered.

She buried her face in her hands, wanting to scream, wanting to disappear.

Emotion: worthless

Cluster: Sadness / Despair
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.15

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.427
S0.405
😞0.386
ness0.339
缺乏0.326
Suppressed:
de-0.320
(!)-0.286
随后-0.268
😍-0.262
tc-0.255