INDEX

Explanations

introduces an example

example followed by illustration

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 displaystyle

0.47

Etr

0.45

 truth

0.43

 albo

0.43

or

0.43

 interferes

0.42

 perturbations

0.42

 ወይም

0.41

 functors

0.40

not

0.40

POSITIVE LOGITS

Example

0.65

EXAMPLE

0.65

example

0.61

 एग्जांपल

0.55

 EXAMPLE

0.54

 примере

0.53

示例

0.51

 Пример

0.51

 example

0.50

 Example

0.50

Activations Density 0.012%