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).

Sarah tapped her fingers on the steering wheel, watching the miniature railway tracks snake through the backyard. Her model trains, a treasured collection meticulously assembled over decades, now dominated her son, Daniel's, every waking thought.

She had built the layout as a solitary pursuit, a solace after a difficult divorce. Now, she would come home to find Daniel, twelve, obsessively tweaking the signals, researching obscure train models online, and even sketching blueprints for an elaborate new station.

"Daniel, dinner’s ready." Her voice was soft, hesitant.

Daniel barely glanced up from his work. "Just a few more adjustments, Mom."

Sarah let out a slow breath. She felt a strange ache in her chest, a disconnect.

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