INDEX

Explanations

to find / to achieve / to create

This neuron activates on definitional or explanatory copular constructions (especially “means” or “is a …”) that introduce a definition or description.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

<unused467>

0.93

<unused606>

0.91

<unused2156>

0.91

<unused2168>

0.91

<unused415>

0.90

<unused2152>

0.88

<unused276>

0.87

<unused613>

0.86

<unused669>

0.85

 intricacies

0.84

POSITIVE LOGITS

0.83

<unused61>

0.76

0.74

<unused62>

0.73

(=

0.73

</code>

0.73

0.71

OP

0.70

}$

0.69

Activations Density 0.061%