INDEX

Explanations

clear association

The neuron is detecting blank‐line or whitespace‐only tokens (i.e. large runs of newlines/spaces).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Squirrel

0.88

Tre

0.82

 Juli

0.81

 Coconut

0.80

0.79

 Este

0.77

Fot

0.77

 Compost

0.76

MMM

0.76

Sun

0.75

POSITIVE LOGITS

 систе

0.77

рг

0.75

зик

0.74

ž

0.74

кг

0.70

ę

0.70

žena

0.69

 пше

0.68

UTION

0.67

 accelerometer

0.67

Activations Density 0.000%