INDEX

Explanations

bracketed numbers

The neuron activates on in‐text reference or footnote markers—that is, bracketed or superscripted citation numbers.

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

in

-1.20

 first

-1.13

 from

-1.11

 most

-1.09

You

-1.07

 року

-1.07

 almost

-1.00

 making

-0.99

but

-0.95

 mostly

-0.94

POSITIVE LOGITS

AWESOME

1.17

 sám

1.15

wiſe

1.14

1.13

 theſe

1.11

Öl

1.10

1.05

 сих

1.04

 paraître

1.03

Activations Density 0.014%