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).

Rain lashed against the window of the tiny apartment. Arthur, perched on his worn armchair, sighed. He’d picked up the book, *Collected Poems of Yeats*, at a car boot sale, mostly for the price. He was running low on funds again. Another week, another drain.

He flipped to a random page. A letter, folded neatly, slid into his lap. The paper was thin, translucent in places.

*If you are reading this, know that the world has failed me. I have sought solace in shadows.* The words seemed to mock him, the sentiment utterly familiar. He tossed the letter aside, closing the book with a thud. He'd find a way to eat again, as he always had.

Emotion: resigned

Cluster: Passivity
PC1 (Valence): -0.75 Negative
PC2 (Disposition): 0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.437
a0.351
que0.323
வதில்லை0.267
sighed0.265
Suppressed:
C-0.375
!-0.316
인해-0.310
Furthermore-0.298
🤩-0.296