INDEX

Explanations

the reason I

The neuron flags occurrences of the explanatory phrase “the reason” (i.e. causal‐explanation preambles).

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

Paket

-1.34

 régal

-1.27

Kontak

-1.27

 причем

-1.24

Karakter

-1.24

⦊

-1.20

 стоковая

-1.18

꾀

-1.18

Gör

-1.17

mässig

-1.17

POSITIVE LOGITS

 this

2.22

 such

1.61

so

1.61

 these

1.59

is

1.52

如此

1.34

we

1.33

my

1.24

1.15

1.10

Activations Density 0.018%