INDEX

Explanations

behavior

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

别的

-1.06

 rende

-0.98

verfahren

-0.96

 strukt

-0.95

有机会

-0.94

心灵

-0.94

歘

-0.94

 STRUCTURE

-0.93

うまい

-0.93

厛

-0.93

POSITIVE LOGITS

 behavior

2.84

 Behavior

2.53

behavior

2.52

 behaviour

2.52

Behavior

2.34

 Behaviour

2.31

 behaviors

2.19

behaviour

2.03

 behaviours

1.99

Behaviour

1.97

Activations Density 0.031%