INDEX

Explanations

specifications or defining attributes

The neuron consistently lights up on formal technical and definitional statements—i.e. sentences laying out specifications, methods, or definitions in a “the … of …” style.

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

 provenientes

-1.48

 trasero

-1.47

 разных

-1.45

 hermosos

-1.44

Understanding

-1.43

liśmy

-1.39

idelberg

-1.37

 zrozum

-1.34

They

-1.34

 same

-1.30

POSITIVE LOGITS

for

3.17

 with

2.05

 when

1.49

 using

1.42

1.41

 should

1.38

者を

1.37

 will

1.33

 without

1.33

his

1.31

Activations Density 0.159%