INDEX

Explanations

handling or using data

The neuron strongly activates on very common “function” words (e.g. of, the, any) and other high‐frequency English tokens.

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

-1.63

an

-1.02

件事

-1.00

one

-0.97

的一个

-0.96

 acolo

-0.92

 wielu

-0.84

>(&

-0.82

 another

-0.82

TH

-0.82

POSITIVE LOGITS

 this

1.90

the

1.75

 этой

1.52

 این

1.44

 these

1.38

 эту

1.37

この

1.34

 that

1.32

这份

1.29

这些

1.27

Activations Density 0.155%