INDEX

Explanations

Python code and English text

The neuron strongly activates on numeric tokens (integer literals) in code snippets.

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

tér

-0.84

游戏特色

-0.77

 igen

-0.75

ambut

-0.74

 تشکر

-0.73

 encer

-0.73

游戏介绍

-0.73

 ус

-0.73

 betek

-0.72

 ярко

-0.72

POSITIVE LOGITS

elif

1.05

[::-

1.02

 heapq

0.92

bise

0.90

 competi

0.89

ſi

0.88

tilia

0.87

を見ながら

0.87

 prate

0.84

-};

0.83

Activations Density 0.010%