INDEX

Explanations

you probably know/noticed/will

The neuron specializes in spotting hedge words that express likelihood or uncertainty (e.g., “probably,” “likely,” etc.).

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

but

-2.50

so

-1.87

may

-1.63

 Such

-1.62

 Currently

-1.57

 this

-1.55

 This

-1.54

 Sometimes

-1.52

<i>

-1.50

and

-1.48

POSITIVE LOGITS

 Zoltán

1.54

 bezw

1.53

 fournisseur

1.51

 ulicy

1.46

zwi

1.45

黝

1.44

鐿

1.41

 делает

1.38

 тоже

1.38

 capitaine

1.35

Activations Density 0.016%