INDEX

Explanations

away from

The neuron strongly activates on the token “away,” effectively detecting instances of the word “away.”

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

</b>

-2.00

 ouvido

-1.99

 voltar

-1.98

𖥧

-1.95

aksikan

-1.93

んですけど

-1.83

我这

-1.82

</i>

-1.81

 rústica

-1.80

我怎么

-1.72

POSITIVE LOGITS

，

2.13

。

2.02

1.91

除了

1.88

如

1.79

他

1.78

「

1.72

``

1.69

*/

1.69

DebuggerNonUser

1.66

Activations Density 0.013%