INDEX

Explanations

cold water

The neuron activates on words and phrases describing low temperatures or cold‐related concepts (e.g. “cold,” “ice,” “chilly,” “cooling”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

اں

-1.68

溪

-1.58



-1.51

 heisst

-1.48

 Dichtung

-1.47

棪

-1.46

 Effluent

-1.46

-1.42

 esquerdo

-1.42

ственного

-1.40

POSITIVE LOGITS

but

1.72

if

1.67

cold

1.61

脔

1.52

 cold

1.52

↵↵↵

1.52

“……”

1.48

 quieran

1.46

 ноги

1.44

责任编辑

1.43

Activations Density 0.066%