INDEX

Explanations

state and its context

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

 douceur

-0.88

込んで

-0.77

 devoir

-0.75

トイレ

-0.72

込んだ

-0.71

 شهرستان

-0.71

َي

-0.71

 davvero

-0.69

טרה

-0.68

siery

-0.68

POSITIVE LOGITS

 state

5.13

 states

4.03

 State

3.94

state

3.75

State

3.47

STATE

3.23

 estado

3.22

 STATE

3.16

状态

3.13

states

3.03

Activations Density 0.055%