INDEX

Explanations

HTML tags and structured text

The neuron activates on XML/HTML element names (i.e. the tag identifiers like “remarks,” “title,” “description,” “p,” etc.) inside angle brackets.

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

]}$

-1.73

%;

-1.66



-1.63

ться

-1.56

できました

-1.55

áticos

-1.55

 felicitación

-1.54

 если

-1.52

YNAMIC

-1.48

onné

-1.48

POSITIVE LOGITS

2.05

as

1.82

 they

1.76

 that

1.74

,,

1.71

↵↵

1.69

for

1.66

啖

1.63

ing

1.58

ul

1.56

Activations Density 0.000%