INDEX

Explanations

section identifiers and contexts

The neuron fires on the literal token “section,” picking out uses of “section” (often followed by numbers or attributes) in code or document headings.

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

actie

-1.92

鲞

-1.81

もありました

-1.80

 créé

-1.74

脸

-1.74

￪

-1.73

 exercer

-1.73

嘥

-1.73

苤

-1.66

 formatos

-1.65

POSITIVE LOGITS

↵

2.83

}{

2.22

</em>

2.11

 continúas

2.06

for

2.06

1.98

section

1.98

</sup>

1.87

1.85

その他の

1.84

Activations Density 0.031%