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).

She found the card nestled amongst bills and junk mail. "Happy Birthday!" it exclaimed in cheerful block letters. Her stomach twisted. The handwriting was unfamiliar, yet a phantom memory tugged at her. Hope, quick and bright, flickered then died.

The kitchen clock ticked, each second a hammer blow against the silence of the empty house. She picked at a loose thread on her sweater, her fingers trembling. The phone sat silent on the counter, a cold, unresponsive monolith. She used to call her sister, but not anymore.

She poured a glass of water, the liquid sloshing over the rim. She should be used to this. Lonely, isolated, unsure. But she wasn't. The card lay forgotten on the counter, a paper reminder of what she didn’t have. She felt a growing, insistent ache in her chest.

Emotion: dependent

Cluster: Shame / Guilt
PC1 (Valence): -1.05 Negative
PC2 (Disposition): 0.69

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.489
缺乏0.395
even0.310
either0.301
'0.297
Suppressed:
笑道-0.309
不久-0.293
一脸-0.263
затем-0.263
庆祝-0.258