INDEX

Explanations

The neuron detects occurrences of “learn” (including “learned,” “learning,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tar

1.34

nél

1.27

tal

1.24

barang

1.24

nál

1.18

că

1.16

donald

1.16

)="

1.16

tik

1.13

INATION

1.13

POSITIVE LOGITS

ية

1.06

悉

1.06

는데

1.05

던

1.04

 firsthand

1.02

на

0.97

 tentang

0.97

게

0.97

 về

0.93

dai

0.91

Activations Density 0.075%