INDEX

Explanations

it is estimated that

The neuron detects numeric expressions (sequences of digits) in the text, especially those giving statistical estimates or counts.

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

噚

-1.60

 What

-1.55

 Seperti

-1.51

 articolo

-1.49

。

-1.47

\&

-1.42

 bocetos

-1.41

}-${

-1.41

 Things

-1.41

all

-1.41

POSITIVE LOGITS

 that

1.73

1.61

時点で

1.58

イケ

1.53

讓你

1.45

对你

1.41

 gdyby

1.37

自体は

1.35

にしても

1.34

),

1.34

Activations Density 0.030%