INDEX

Explanations

reduce or prevent negative outcomes

The neuron fires on technical-purpose verbs that describe what something does or is meant to do (e.g. distribute, prevent, provide, reinforce).

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

by

-1.68

が見える

-1.39

 administrativa

-1.24

鯵

-1.22

の人に

-1.22

preced

-1.17

 like

-1.15

 American

-1.13

exer

-1.13

気がする

-1.13

POSITIVE LOGITS

if

1.42

Jeśli

1.30

setResizable

1.28

ここ

1.27

สำหรับ

1.27

 Meningkatkan

1.23

大部分

1.20

tividade

1.19

Biografia

1.16

 primeiras

1.16

Activations Density 0.198%