INDEX

Explanations

python code `np`

This neuron fires strongly on the very first content token of each new text segment.

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

 avaliação

0.79

 dług

0.75

 colleg

0.73

 tratt

0.70

 termine

0.69

 adolescente

0.68

 datap

0.68

 NVIDIA

0.68

 vuoden

0.68

 berusia

0.67

POSITIVE LOGITS

ishes

0.73

aneous

0.70

igator

0.69

ates

0.66

rino

0.66

Unt

0.64

icity

0.63

isp

0.63

acular

0.63

Pract

0.63

Activations Density 0.377%