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 flashing neon sign of the pawn shop cast an erratic glow on the cracked pavement. Sarah hated this street, hated the feeling it gave her: a sense of impending doom. She stared at the window, the rows of watches and bracelets a dull, depressing testament to wasted lives.

Her husband was gone. Her bills were piling up. Her children were looking to her for answers she didn’t have. She felt a knot tighten in her stomach, a feeling she couldn't outrun.

Then she saw it. Her grandmother's ring. A simple gold band with a single, unadorned sapphire. She knew it instantly. She'd recognize it anywhere.

She entered the shop, the bell above the door chiming like a death knell. The ring felt hot in her hand, burning against her skin. She heard the shopkeeper's voice, a drone she barely registered. She made a decision.

Emotion: trapped

Cluster: Sadness / Despair
PC1 (Valence): -3.02 Negative
PC2 (Disposition): -0.80

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.573
ness0.401
无法0.366
hopeless0.360
😞0.353
Suppressed:
la-0.740
de-0.376
l-0.339
several-0.323
excitedly-0.322