INDEX

Explanations

Q: questions and labels

The neuron fires primarily on standalone numeric tokens (i.e. sequences of digits) within code or markup.

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

 injections

-0.86

 SUSTAIN

-0.82

 activer

-0.82

 جست

-0.81

 isinstance

-0.81

 המע

-0.81

angerschaft

-0.80

 embry

-0.79

Eks

-0.77

UGG

-0.76

POSITIVE LOGITS

<-

0.78

 sentarse

0.77

 wcześniej

0.77

 Appalach

0.75

Qu

0.74

What

0.73

 unterschied

0.73

Tra

0.73

mx

0.71

 había

0.71

Activations Density 0.006%