A dull ache throbbed behind David’s eyes as he scrolled through the online food blog. It was late, he was curled on the couch, laptop precariously balanced on his knees. He’d spent the whole day hunched over spreadsheets, the numbers blurring into an indecipherable mess. He stumbled upon an article about the "Golden Glow Cake," a dessert he’d perfected years ago, a recipe he'd shared with a friend. The friend, now a successful baker, was being hailed as a culinary genius.

He let out a long breath, pushing the glasses up his nose. The article gushed over its ethereal lightness and subtle citrus notes. He knew it was his recipe. His hands felt heavy, almost numb, as he closed the laptop. The room seemed suddenly too warm, the silence pressing down on him.

Emotion: tired

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.92 Negative
PC2 (Disposition): -1.00

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
weary0.254
fatigue0.249
😴0.238
sighed0.237
barely0.230
Suppressed:
C-0.268
because-0.225
B-0.219
🤟-0.217
భి-0.214