INDEX

Explanations

$Q$ or `Use` or `<!`

This neuron fires on occurrences of square‐bracket delimiters (e.g. “[” and “]”) in mathematical expressions, such as indexing or expectation notation.

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

最初は

-1.63

いかがでしょうか

-1.56

他有

-1.55

 использовать

-1.55

 genauer

-1.54

つは

-1.54

她也

-1.53

元々

-1.52

 Когда

-1.49

POSITIVE LOGITS

ět

1.71

ÑOS

1.63

 花卉

1.63

 男士

1.61

 dezelve

1.60

ússia

1.54

นี้

1.53

鉝

1.51

Ꭽ

1.50

 llam

1.48

Activations Density 0.004%