INDEX

Explanations

cast / casting

The neuron fires whenever it sees the letter sequence “cast” embedded in a token (e.g. in words like “castrol,” “castle,” “cache,” “showcase,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ynucle

0.61

్వర

0.61

 ফিল

0.61

ு

0.56

ख्यात

0.55

erche

0.55

tri

0.53

พย์

0.53

駕

0.53

Три

0.53

POSITIVE LOGITS

 Cast

1.39

 cast

1.38

Cast

1.31

 casts

1.23

cast

1.19

 casting

1.19

CAST

1.19

 CAST

1.16

 Casting

1.16

 castor

1.15

Activations Density 0.115%