INDEX

Explanations

enable, disable, select states

The neuron activates on uppercase hardware‐register field names and constants (e.g. ENABLE, INTEN, CLR).

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

怎样的

-1.55

ch

-1.49

<0xCE>

-1.48

⢮

-1.45

儅

-1.44

 kurzer

-1.41

“

-1.41

 possibly

-1.40

 about

-1.39

how

-1.39

POSITIVE LOGITS

}$.

1.74

 према

1.63

ים

1.59

and

1.59

𖧵

1.58

媖

1.54

𓆜

1.52

 для

1.48

zboi

1.48

for

1.45

Activations Density 0.008%