INDEX

Explanations

crosses through

The neuron detects verbs that describe a road or river traversing, crossing, or running through a location.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 through

-1.32

 diferite

-1.30

as

-1.27

同上

-1.27

utilizzo

-1.24

sa

-1.24

 with

-1.23

 whofe

-1.23

文中

-1.22

(_.

-1.22

POSITIVE LOGITS

gesamt

1.27

つづく

1.25

どうしよう

1.25

你不要

1.24

 visites

1.24

to

1.24

 itself

1.23

孱

1.23

Aucun

1.21

帰って

1.21

Activations Density 0.076%