INDEX

Explanations

programming code and chess

The neuron is detecting numeric literal constants in code (e.g. numbers like 198, 159, 173, etc.).

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

-1.73

add

-1.48

 использова

-1.45

of

-1.44

 unseren

-1.41

见过

-1.35

op

-1.32

太多了

-1.31

あげて

-1.30

钨

-1.30

POSITIVE LOGITS

and

1.85

1.63

 ajedrez

1.57

âncias

1.52

 xadrez

1.51

'},

1.46

 qualities

1.45

🫅

1.45

boj

1.45

thief

1.44

Activations Density 0.016%