INDEX

Explanations

i

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 StringTokenizer

-0.72

es

-0.65

InputBorder

-0.54

 internationales

-0.53

tling

-0.53

sind

-0.51

 fibrillation

-0.51

 vectorielle

-0.50

 condensation

-0.49

 discharging

-0.49

POSITIVE LOGITS

<bos>

0.80

was

0.74

principalColumn

0.71

 likes

0.66

 estekak

0.66

 makes

0.65

 thinks

0.65

 antaranya

0.63

 would

0.62

Activations Density 0.191%