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 attic air hung thick and dusty, catching in Eleanor’s throat. A trunk, crammed with moth-eaten treasures, sat beneath a cobweb-draped chandelier. Inside, amongst faded photographs and forgotten toys, lay a leather-bound journal. It was her mother’s handwriting, neat and familiar. Flipping it open, she found a handwritten list: "See the Northern Lights. Learn to bake a perfect croissant. Visit the Taj Mahal." A knot tightened in Eleanor’s stomach. She ran a hand through her hair, the volume of her mother's unlived life pressing down on her.

The entries after the list were fragmented, half-finished thoughts, doodles of birds taking flight, and dried pressed flowers. Eleanor felt a strange disorientation, a disconnect between the vibrant woman she remembered and these quiet, unfulfilled dreams. She sank onto a nearby crate, the wood digging into her. What had her mother been hiding?

Emotion: puzzled

Cluster: Surprise / Confusion
PC1 (Valence): -1.17 Negative
PC2 (Disposition): -0.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.453
🤔0.421
🤔0.402
🧐0.328
a0.294
Suppressed:
ecstatic-0.265
gladly-0.238
swiftly-0.237
alleviate-0.234
wholeheartedly-0.232