INDEX

Explanations

unlimited proxy

The neuron activates on standalone numeric version tokens (e.g. “V3,” “d3.js,” “ipython‐19.5”) in code or API/library references.

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

ęty

-1.58

cerer

-1.41

rakech

-1.38

 commandés

-1.37

 protéger

-1.32

anleitung

-1.32

TypedDataSet

-1.30

rungs

-1.29

urther

-1.28

POSITIVE LOGITS

of

2.58

 from

1.66

 what

1.56

 should

1.45

 provides

1.41

What

1.34

 gives

1.32

 היתה

1.31

 could

1.30

 puts

1.28

Activations Density 0.000%