INDEX

Explanations

code, buyers, How, Q

The neuron fires on technical/documentation keywords—things like “batch,” “input,” “data,” etc.

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

for

-1.56

in

-1.55

by

-1.46

if

-1.42

to

-1.36

 after

-1.36

as

-1.31

so

-1.28

on

-1.27

 with

-1.20

POSITIVE LOGITS

 relâche

1.27

久し

1.25

voyez

1.23

 prega

1.16

 appartiennent

1.16

 brachten

1.14

の準備

1.14

 mélanger

1.13

 territo

1.12

 délais

1.12

Activations Density 0.002%