INDEX

Explanations

will be

This neuron detects the presence of modal auxiliary verbs expressing future or conditional tense (e.g., “will,” “would”).

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

 thăm

1.11

Jähr

1.05

 Dung

1.03

 Coleoptera

0.99

 αριθ

0.98

 sẽ

0.97

 Arzneim

0.96

 would

0.95

 finement

0.94

arxiv

0.94

POSITIVE LOGITS

1.89

oughby

1.32

 sicuramente

1.24

 tetapi

1.21

1.16

1.14

 certamente

1.12

lás

1.10

lly

1.09

Activations Density 0.696%