INDEX

Explanations

probability of sequence without replacement

The neuron detects LaTeX‐style mathematical notation (e.g. $$, \begin, array, \\) and formula tokens.

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

^(@

-2.02

珦

-2.00

 ſer

-1.83

 ſtand

-1.73

ехали

-1.70

裢

-1.70

 และ

-1.69

eſ

-1.66

 miſ

-1.64

==""){

-1.63

POSITIVE LOGITS

 during

2.09

 with

2.02

 particularly

1.80

 good

1.73

but

1.67

 these

1.66

1.65

 although

1.64

 used

1.63

 provide

1.63

Activations Density 0.076%