INDEX

Explanations

The neuron is detecting numeric tokens (e.g. indices, counts, and other number strings).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 both

0.56

Trigger

0.55

OFF

0.55

痪

0.54

 दोनों

0.53

 새로

0.52

both

0.52

 determinada

0.52

Naw

0.52

 וח

0.51

POSITIVE LOGITS

,...,

1.60

 ...,

1.49

+...+

1.37

...,

1.36

 …,

1.31

…,

1.15

 ..."

1.14

...

1.13

......

1.13

.......

1.13

Activations Density 0.019%