INDEX

Explanations

directly to or from

The neuron fires on tokens expressing directness (e.g. “direct,” “directly”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 autre

-1.75

("

-1.69

-1.47

-1.44

-1.41

</u>

-1.41

-1.39

готовки

-1.37

-1.36

POSITIVE LOGITS

 such

2.02

 then

1.73

婓

1.72

 takiej

1.62

 pazi

1.62

 Русский

1.59

 proposte

1.57

</h2>

1.52

 trucos

1.52

izontal

1.50

Activations Density 0.042%