INDEX

Explanations

Hunter Oaks, Though he, Variations in

The neuron strongly activates on standalone numeric tokens (digits or digit sequences) in code, such as version numbers or numeric identifiers.

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

</h1>

-1.49

Étymologie

-1.45

Вариант

-1.34

Publié

-1.27

Клю

-1.27

Beskrivelse

-1.27

</strong>

-1.25

-1.24

 will

-1.23

Каждый

-1.23

POSITIVE LOGITS

 about

1.90

 more

1.68

1.56

archical

1.55

the

1.51

of

1.45

 that

1.43

⁶

1.37

(()=>{

1.37

Mutagenicity

1.35

Activations Density 0.004%