INDEX

Explanations

loops and array dimensions

The neuron fires on identifiers that name or refer to a container’s length/size (e.g. N, str.length).

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

に来

-1.42

づけ

-1.38

citroen

-1.34

 vinilo

-1.34

為に

-1.32

 zelve

-1.30

carrot

-1.30

 cinturón

-1.29

 lavadora

-1.27

joven

-1.25

POSITIVE LOGITS

これは

1.38

1.27

に

1.26

は

1.26

そして

1.25

と

1.22

ですが

1.21

を

1.20

そのため

1.20

で

1.19

Activations Density 0.002%