INDEX

Explanations

specific states or descriptions

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-1.89

 then

-1.78

 always

-1.77

 have

-1.57

 another

-1.54

 suggests

-1.51

อื่น

-1.48

 دیگری

-1.48

อีก

-1.44

 know

-1.43

POSITIVE LOGITS

 даже

1.82

 nawet

1.47

</i>

1.45

 bahkan

1.41

<td>

1.38

ov

1.34

 mengumumkan

1.28

 厂家

1.22

 właśnie

1.21

 навіть

1.21

Activations Density 0.156%