INDEX

Explanations

Q: and Stronger

This neuron fires strongly on the attribute name “is” in XML/HTML‐style tags (e.g. `<template is="…">`).

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

mathrm

-1.78

 valo

-1.76

他的

-1.75

 думал

-1.74

 декабря

-1.68

亊

-1.68

～

-1.59

╿

-1.56

 namun

-1.52

It

-1.52

POSITIVE LOGITS

!"

1.57

惴

1.56

 superb

1.50

("

1.49

 verbess

1.48

我們

1.47

1.45

 tych

1.40

 estos

1.39

噤

1.39

Activations Density 0.000%