INDEX

Explanations

followed by sequence or continuation

This neuron activates on numeric tokens—sequences of digits such as years, timestamps, or other numbers.

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

 from

-1.18

on

-1.16

 their

-1.09

 τὴν

-1.08

 that

-1.07

now

-1.05

if

-1.04

 with

-0.99

 pirm

-0.97

ఒ

-0.97

POSITIVE LOGITS

 prosegu

1.45

 then

1.41

 followed

1.34

 フー

1.24

 conferma

1.20

之后

1.17

之後

1.13

 puis

1.12

然后

1.06

ciled

1.06

Activations Density 0.004%