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 vibrant decorations, the colourful balloons, the carefully prepared snacks, all seemed to leer at Olivia. A single, short text message: "Things have changed. No party." And that was it.

Sheโ€™d spent the morning cleaning, baking, trying to create an atmosphere of joy, a vessel for connection. Now, the apartment was a tomb of unrealized expectations. The radio played a jaunty tune, but the music felt like a cruel joke.

She wandered through the rooms, touching the surfaces, the objects seeming to hum with a silent disapproval. She felt herself shrinking, becoming invisible, the air around her thick with unspoken judgements.

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
๐Ÿ˜ž0.386
ness0.339
็ผบไน0.326
Suppressed:
de-0.320
(!)-0.286
้šๅŽ-0.268
๐Ÿ˜-0.262
tc-0.255