INDEX

Explanations

schedule related contexts

The neuron fires on highly technical or structured‐data tokens—things like code or log keywords, variable names, and other specialized tokens (e.g. “schedule” in a log, “m^” in code, “initialize” in a model definition).

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

might

-1.71

eding

-1.70

ſo

-1.61

 Mungkin

-1.59

 requiere

-1.58

 Siapa

-1.55

 REALLY

-1.55

 kust

-1.53

believe

-1.52

could

-1.50

POSITIVE LOGITS

 then

2.31

at

2.22

 with

2.16

now

2.03

 using

1.95

for

1.89

 since

1.86

 during

1.70

 after

1.66

 should

1.63

Activations Density 0.000%