INDEX

Explanations

a.m. and p.m

The neuron detects the single-letter “a” or “p” used as part of time‐of‐day markers (as in “a.m.” and “p.m.”).

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

</strong>

-2.78

 when

-2.45

).

-2.41

,’

-2.19

 what

-2.16

-2.14

if

-2.13

il

-2.11

but

-2.08

 республики

-2.08

POSITIVE LOGITS

땃

2.20

釙

2.17



1.98



1.97

拶

1.89

 susc

1.87

 Parlement

1.86

ᜀ

1.84

反而

1.84

 ducha

1.84

Activations Density 0.009%