INDEX

Explanations

community care and services

The neuron fires on occurrences of the word “community,” especially when it’s part of an organization or service name.

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

seriously

-0.75

бора

-0.73

yt

-0.72

专题

-0.70

暮らし

-0.69

รส

-0.69

vinfos

-0.69

ontan

-0.69

GLOB

-0.69

ɫ

-0.68

POSITIVE LOGITS

splitlines

0.84

flikt

0.77

 groups

0.75

 asset

0.75

terer

0.72

dead

0.72

釦

0.71

 payback

0.71

 focused

0.71

department

0.70

Activations Density 0.016%