gloomy - Story 54250 | Emotion Vectors in Gemma4-31B

The city lights blurred through the rain-streaked window of the taxi. Emily stared out at the passing scenery, the vibrant hustle of the city a stark contrast to the quiet despair inside her. The phone call. The words, "medical record mix-up." They continued to replay in her mind.

She had been looking forward to meeting her friends at the bar. Now, the thought of social interaction seemed exhausting. She leaned back in her seat, closing her eyes, wishing she could simply disappear.

The bar lights and music, the noise and conversation, would be the opposite of what she needed. She was glad, at least, that she had decided to stay home. The taxi ride felt like a slow descent, a fall into a bottomless pit of concern.

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
`쓸`	0.255

Suppressed:

`la`	-0.432
`l`	-0.419
`de`	-0.398
`B`	-0.294
`!`	-0.259