INDEX

Explanations

discusses or states

The neuron fires on words or tokens that mean “is about” or “deals with” (e.g. Italian tratta, German handelt, Spanish trata, English dealing).

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

禩

-2.08

厹

-2.00

 versátil

-1.98

and

-1.83



-1.82

 воздей

-1.81

 damages

-1.81

ᚺ

-1.79

嚶

-1.78

</td>

-1.75

POSITIVE LOGITS

 There

2.61

 That

2.16

.”

2.09

2.08

 However

1.91

It

1.88

‘

1.84

’

1.82

</strong>

1.80

 Unfortunately

1.80

Activations Density 0.001%