INDEX

Explanations

`(` and `"`

The neuron is selectively activating on the double‐quote characters that delimit string literals (i.e. the “"” marks) in source code.

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

 peux

-1.86

of

-1.75

 aumentando

-1.74

re

-1.67

 existing

-1.66

 aumentar

-1.63

 Weisheit

-1.61

-1.59

ของ

-1.58

إذا

-1.57

POSITIVE LOGITS

↵

1.64

 Многие

1.61

 Даже

1.55

 roasting

1.45

衣装

1.44

Odor

1.43

 hábit

1.43

蹋

1.40

atella

1.39

 Сейчас

1.38

Activations Density 0.002%