INDEX

Explanations

technical configuration parameters

The neuron strongly activates on programming‐or API‐related acronyms and technical identifiers (e.g. “URL,” “IP,” “CSV,” “OAuth,” etc.).

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

pherds

-1.27

 tanques

-1.23

 MÁS

-1.22

 fatiga

-1.22

 bailarina

-1.20

 könig

-1.20

 nagyobb

-1.20

 moletom

-1.17

 bené

-1.16

geführten

-1.14

POSITIVE LOGITS

for

1.90

or

1.56

 that

1.50

 whose

1.41

 which

1.41

 will

1.34

of

1.29

who

1.21

 الذي

1.20

can

1.20

Activations Density 0.018%