INDEX

Explanations

code keywords and library names

The neuron is triggered by code‐related tokens naming libraries, modules, or commands (e.g. import/module names and CLI tool names).

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

 with

-1.69

 they

-1.45

for

-1.34

 their

-1.33

at

-1.32

 those

-1.32

all

-1.26

 some

-1.20

 each

-1.13

 these

-1.10

POSITIVE LOGITS

schuhe

1.38

 débat

1.20

 poils

1.19

Ingredienti

1.16

küche

1.16

Kdo

1.14

влений

1.13

带来了

1.10

Verwaltung

1.09

锝

1.09

Activations Density 0.002%