INDEX

Explanations

code examples or tutorials

The neuron fires on tokens naming sample or example code artifacts (e.g. “sample,” “Sample,” “Example,” “Demo”), i.e. labels for tutorial or demo source files.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

the

-1.74

 because

-1.47

⃣

-1.38

:]

-1.31

 solusi

-1.30

Ǆ

-1.27

itarios

-1.24

 Populations

-1.23

(%)

-1.23

 wanneer

-1.21

POSITIVE LOGITS

↵↵

1.54

1.52

 получила

1.42

та

1.41

re

1.38

 كورة

1.34

which

1.32

Didn

1.30

("

1.30

ness

1.27

Activations Density 0.057%