INDEX

Explanations

at specific values

The neuron activates on numeric measurement tokens (i.e. numbers indicating wavelengths, temperatures, distances, time points, etc.) in scientific text.

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

慟

-1.52

貿易

-1.34

それは

-1.33

弔

-1.32

揶

-1.24

was

-1.21

 Ordens

-1.20

 Nationen

-1.20

 ложка

-1.19

of

-1.11

POSITIVE LOGITS

itern

1.23

__________

1.20

の方も

1.19

に入った

1.14

にの

1.13

[]={

1.13

 Their

1.11

ulit

1.11

 både

1.09

Ric

1.07

Activations Density 0.089%