INDEX

Explanations

links in documentation

This neuron detects the opening square bracket of markdown-style links or reference tags.

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

<bos>

-1.82

if

-1.48

-1.35

 only

-1.32

for

-1.30

 corrobor

-1.29

 with

-1.26

-1.24

商店街

-1.21

POSITIVE LOGITS

 wohn

1.33

psies

1.32

 berichtet

1.27

#\

1.27

OO

1.27

umbers

1.24

1.23

ᴘ

1.22

 residuos

1.21

 imprimer

1.21

Activations Density 0.052%