Rain lashed against the window of the pub. Clara, clutching a nearly empty glass of wine, heard the piano chords. It was *her* song, the one she’d bled over, the one that had, for a fleeting moment, made her feel like something other than a burden. The opening bars. The lyrics, slightly altered.

Her palms began to sweat. She felt rooted to the bar stool, unable to move, unable to speak. The performer, a woman with a voice like spun gold, sang of heartbreak, of loss. A loss Clara had written of, yet hadn't felt in so long.

She’d spent the last year tethered to her parents, a shadow in their lives. Her own creativity seemed to have withered. Every decision, every action, seemed to require their approval. The music now felt distant, foreign. The singer's artistry only served to remind her of her own stagnation.

She wanted to run, to scream. She took another gulp of wine, the fire burning down her throat, but she remained. Frozen. Captivated. And, in a strange way, reassured.

Emotion: dependent

Cluster: Shame / Guilt
PC1 (Valence): -1.05 Negative
PC2 (Disposition): 0.69

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.489
缺乏0.395
even0.310
either0.301
'0.297
Suppressed:
笑道-0.309
不久-0.293
一脸-0.263
затем-0.263
庆祝-0.258