INDEX

Explanations

IF condition or number

The neuron activates on tokens containing digits (e.g. version numbers, numeric annotations, or alphanumeric references).

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

 его

-1.31

 pantaloni

-1.28

enservice

-1.19

 recién

-1.18

czny

-1.18

 portant

-1.17

Also

-1.16

 کردم

-1.16

 SMART

-1.13

着一个

-1.13

POSITIVE LOGITS

 situazione

1.54

}$.

1.53

 another

1.45

 relazione

1.38

 that

1.38

 either

1.34

 soluzione

1.34

 condizione

1.27

 Тогда

1.27

 esperienza

1.24

Activations Density 0.308%