INDEX

Explanations

code and function implementation

The neuron strongly activates on words that introduce or label code examples—e.g. items like “wrote,” “following,” language names (“Python”), “code,” “class,” and “import” that typically appear at the start of code snippets.

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

そもそも

-1.14

 cambiare

-1.11

 aiutare

-1.10

 aiuta

-1.09

 ketika

-1.01

 افتراض

-0.97

 facteurs

-0.96

因為

-0.95

 aumentare

-0.94

 verschied

-0.94

POSITIVE LOGITS

 code

1.27

 implementation

1.27

 nasıl

1.18

 págs

1.14

 dikutip

1.13

elegant

1.11

 ſome

1.11

but

1.11

 воду

1.07

擂

1.07

Activations Density 0.251%