INDEX

Explanations

semicolon terminated code statements

The neuron is keyed to the appearance of the token “Activity” (e.g. in Android Activity class declarations).

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

 Since

-0.94

 AssemblyCulture

-0.87

 Pří

-0.83

拼接

-0.80

 memorable

-0.80

 Berlin

-0.80

mà

-0.79

Br

-0.78

Ni

-0.77

 disparate

-0.77

POSITIVE LOGITS

import

1.40

↵↵↵↵

1.35

↵↵↵↵↵↵

1.18

↵↵↵↵↵

1.10

↵↵↵↵↵↵↵↵↵↵

1.07

↵↵↵↵↵↵↵↵↵

1.05

↵↵↵↵↵↵↵

1.04

public

1.03

↵↵↵↵↵↵↵↵

1.03

Ưu

1.02

Activations Density 0.002%