INDEX

Explanations

The neuron fires most strongly on tokens that are parts of file‐path or URL strings for linked resources (e.g. “.js,” “.css,” and other src/href path segments).

javascriptassetpaths

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

on

-1.38

our

-1.18

 came

-1.15

 there

-1.08

 this

-1.07

確認ください

-1.06

my

-1.06

if

-1.05

one

-1.03

 early

-1.02

POSITIVE LOGITS

توضیحات

1.48

 traitements

1.23

 possibilités

1.22

Oby

1.19

 comportements

1.15

 besoins

1.14

 tillegg

1.12

 actuelles

1.11

مقدمه

1.10

NEED

1.08

Activations Density 0.015%