INDEX

Explanations

connected to something

The neuron consistently fires on main lexical verbs that state relationships or conditions (e.g. “connecting,” “consists,” “live,” “existed,” “contained,” etc.).

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

 opportunità

-0.96

ܕ

-0.94

amitié

-0.93

ァ

-0.87

ﻥ

-0.87

νό

-0.86

拿到

-0.85

åter

-0.85

Recu

-0.84

 rozhod

-0.84

POSITIVE LOGITS

and

1.35

0.93

riente

0.90

cv

0.90

需要

0.89

iotti

0.88

“

0.86

telah

0.85

“

0.85

 Apos

0.83

Activations Density 0.102%