INDEX

Explanations

List or array of strings

This neuron detects the angle‐bracket tokens (“<” and “>”) that mark generic type parameters 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

For

-2.38

-2.22

-2.11

切割

-2.03

 troubling

-1.95

</h2>

-1.86

があります

-1.79

Tecnologia

-1.76

-1.74

);

-1.66

POSITIVE LOGITS

2.91

2.42

2.23

orothy

2.19

 أو

2.06

ientras

2.03

 that

2.03

2.02

 allora

2.00

 และ

1.99

Activations Density 0.019%