"Just five more pages," Maya mumbled, her voice thick and slow as she fought to stay upright. The late afternoon sun streamed through her window, illuminating dust motes that danced in the air. She’d been up since dawn, chasing the elusive deadline for her article. The subject: the works of Eleanor Vance, her favourite poet.

Her eyelids drooped, heavy and uncooperative. She propped her chin on her hand, the weight of her head making her arm ache. The laptop screen flickered, the cursor blinking, mocking her inability to concentrate. She’d promised herself a break after finishing this paragraph.

She’d just saved the file when a frantic email popped up. A friend had sent a link, with a subject line: "Eleanor Vance? Seriously?" Maya clicked the link, her brain feeling like it was packed with cotton. The article was a bombshell. Vance, her beloved Vance, had been lifting entire stanzas, whole poems, from forgotten poets of the Victorian era. The blood drained from Maya's face. The room started to spin. Her favourite poems, built on lies.

Emotion: sleepy

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.86 Negative
PC2 (Disposition): -1.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😴0.318
weary0.310
fatigue0.296
💤0.283
似乎0.280
Suppressed:
B-0.343
C-0.294
because-0.293
భి-0.281
เพราะ-0.277