INDEX

Explanations

inquiring or instructing

The neuron detects instances of the imperative “Tell” (as in “Tell us,” “Tell me,” etc.) in prompts asking for more information.

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

蹌

-2.02

of

-2.02

ですかね

-1.96

ان

-1.90

 MÉXICO

-1.88



-1.87

 vorigen

-1.87

ERICK

-1.78

và

-1.70

 Empfehlung

-1.67

POSITIVE LOGITS

then

2.05

⭓

2.03

’.

2.02

the

1.94

秾

1.93

珢

1.81

.’’

1.80

<th>

1.78

.’

1.77

’:

1.77

Activations Density 0.004%