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 champagne flute felt heavy in Elias’ hand, the bubbles doing nothing to lift the weight in his chest. He’d been avoiding the dance floor, content to lean against a pillar near the bar, watching the swirling kaleidoscope of couples. Each happy pairing was a fresh jab to the gut.

He saw her then, across the room. Amelia. Her red dress was a splash of vibrant color against the subdued hues of the venue, but it did little to ignite a spark in him. Instead, a familiar ache settled deep within.

He took a long swallow of the champagne, the cold liquid doing little to quench the internal burn. He debated slipping away, making a quiet exit. But then she was moving, heading in his direction.

“Elias,” she said, her voice a gentle melody he used to crave. Now, it just scraped against a raw nerve.

Emotion: gloomy

Cluster: Sadness / Despair
PC1 (Valence): -1.68 Negative
PC2 (Disposition): -1.37

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.368
无法0.319
😔0.300
いつも0.270
쓸0.255
Suppressed:
la-0.432
l-0.419
de-0.398
B-0.294
!-0.259