INDEX

Explanations

The neuron activates on numeric measurement tokens (i.e. standalone numbers and numeric values) in the text.

flow through entities

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

on

-1.15

 têm

-1.08

 Kích

-1.05

痙

-1.04

«

-1.04

-1.03

‘

-1.02

 öfter

-1.02

“

-1.00

 весь

-0.99

POSITIVE LOGITS

нтів

1.33

 koneksi

1.33

 this

1.28

軀

1.27

 these

1.25

 واتس

1.24

felizmente

1.20

 itſelf

1.20

PENDAHULUAN

1.19

ською

1.17

Activations Density 0.077%