INDEX

Explanations

hexadecimal identifiers

This neuron detects hexadecimal strings (e.g. long runs of 0–9 and A–F characters) in code.

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

 ícone

-1.16

 ícones

-1.14

 bellissimo

-1.13

現象

-1.11

 leão

-1.10

 anjo

-1.09

 tranquillo

-1.09

 okolí

-1.08

 касается

-1.07

必須

-1.05

POSITIVE LOGITS

the

1.36

rsquo

1.30

viembre

1.24

}\}

1.22

 encantan

1.21

 различни

1.14

imab

1.12

żeń

1.12

ysteine

1.12

 Films

1.12

Activations Density 0.012%