INDEX

Explanations

equation solutions

The neuron activates on mentions of “solution” (and its verbal forms like “solve” or “solving”) in mathematical contexts.

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

-1.77

搽

-1.19

-1.17

ang

-1.16

佝

-1.15

üş

-1.14

 correndo

-1.13

 passando

-1.13

まわ

-1.13

 mögliche

-1.12

POSITIVE LOGITS

of

3.53

for

1.87

 with

1.55

banyak

1.38

 substantially

1.32

 ottenuto

1.30

 without

1.30

🎧

1.25

 alſo

1.25

uniqlo

1.24

Activations Density 0.019%