INDEX

Explanations

finding it

The neuron strongly responds to occurrences of the verb “find.”

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

ج

1.88

ه

1.76

ش

1.58

 निभाने

1.44

торы

1.34

ል

1.34

cL

1.29

ことになる

1.29

بة

1.28

ção

1.27

POSITIVE LOGITS

 fault

1.39

ividual

1.28

ана

1.22

 Nemo

1.22

amental

1.19

 した

1.19

ੇ

1.19

 primi

1.13

్

1.12

ाइंड

1.11

Activations Density 0.128%