INDEX

Explanations

learning process and goals

The neuron chiefly detects occurrences of the word “learning.”

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

↵

-2.66

 "/")

-2.14

'./

-1.96

 这个

-1.91

 normale

-1.86

玧

-1.82

 "-";

-1.81

 怎么

-1.81

从

-1.80

这

-1.78

POSITIVE LOGITS

.,"

2.06

 Lernen

2.02

鷥

1.84

.'"

1.83

 transferred

1.83

⏑

1.80

を書

1.80

 intently

1.79

⮑

1.77

 aprendizaje

1.75

Activations Density 0.033%