INDEX

Explanations

graphics API resource management

The neuron activates on numeric tokens—e.g. integer literals and standalone numbers.

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

转换为

-0.91

Server

-0.82

rmi

-0.79

 Pittsburgh

-0.78

Opened

-0.75

 Mediterranean

-0.75

Save

-0.75

jSON

-0.75

硏

-0.73

unting

-0.72

POSITIVE LOGITS

 swap

1.00

 Swap

0.90

kových

0.86

swap

0.85

 RECRE

0.84

 Instance

0.84

 プール

0.84

cleanup

0.82

 recreate

0.76

 Recre

0.76

Activations Density 0.004%