INDEX

Explanations

your data, new york, can help

This neuron detects words in headlines or section titles, especially title-cased or prominent words in article headings.

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

 広場

-1.34

himself

-1.29

lorn

-1.27

zey

-1.27

opie

-1.25

 graines

-1.23

 tekno

-1.23

 opini

-1.22

lled

-1.21

lösung

-1.21

POSITIVE LOGITS

for

1.97

 their

1.81

 which

1.53

 Their

1.52

 with

1.42

Why

1.41

 where

1.40

 Unlike

1.34

 from

1.33

 whose

1.31

Activations Density 0.294%