INDEX

Explanations

testing and expectations

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

if

-4.09

 если

-2.38

If

-2.06

if

-1.91

 якщо

-1.84

If

-1.79

 jeśli

-1.77

 ако

-1.74

 nếu

-1.74

 dacă

-1.73

POSITIVE LOGITS

 EXPECT

1.48

 expect

1.47

expect

1.44

 expecting

1.42

 expectation

1.40

EXPECT

1.38

assertEquals

1.33

 quando

1.28

 Expect

1.27

 expectations

1.26

Activations Density 0.022%