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 concrete felt cold beneath Maya’s knees. She’d been working as a receptionist for years, feeling as though she was fading. The monotony of the phone ringing, the same conversations, the same fluorescent lights…it was soul-sapping. She was pulling up weeds, her shoulders aching from the repetitive motion, in a small, neglected park she’d chosen to take a shortcut through. She felt the dull weight of her usual mood in her chest, a constant, low thrum.

Her trowel struck something hard, a resistant thump in the soft earth. She unearthed a heavy, rectangular container. A simple metal box. Inside, carefully sealed, were letters, a dried flower, and a small, tarnished silver locket. Each item stirred something in her, a feeling she couldn't name. The letters bore her childish handwriting. A younger version of herself, full of life, stared back from the past.

Emotion: dispirited

Cluster: Sadness / Despair
PC1 (Valence): -1.88 Negative
PC2 (Disposition): -0.98

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.398
S0.348
’0.325
ness0.320
无法0.308
Suppressed:
de-0.297
l-0.256
👏-0.251
sonra-0.240
aud-0.240