INDEX

Explanations

programming keywords and constructs

The neuron fires on programming‐language punctuation and structural tokens (e.g. dots, braces, include directives), in other words it identifies source-code delimiters.

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

 dezelve

-1.15

菒

-1.02

 memória

-0.98

DAVID

-0.97



-0.96

 leſs

-0.95

┹

-0.94

᠁

-0.93

сторінку

-0.93



-0.93

POSITIVE LOGITS

or

1.02

Denote

0.81

!!!!

0.77

In

0.76

Are

0.75

Upstairs

0.75

ựu

0.75

!!!!!

0.74

Activations Density 0.014%