INDEX

Explanations

fitting

The neuron fires on any occurrence of the root “fit” (e.g. fit, fitted, fitting, fitters) regardless of context.

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

鱻

-2.19

ruvate

-2.08

 nowa

-2.02

沇

-2.02

蜑

-2.00

eau

-1.99

鷓

-1.97

绝望

-1.96

Ꮷ

-1.91

ྥ

-1.89

POSITIVE LOGITS

 When

2.28

 cuando

2.27

 They

2.27

hatian

2.22

」

2.09

 This

2.08

でて

2.06

 quando

2.05

 好看

2.02

 However

1.98

Activations Density 0.014%