INDEX

Explanations

names or entities with subsequent verbs

The neuron fires on short phrases expressing uncertainty, inability, or difficulty (e.g. “Nothing comes to mind,” “sometimes is hard for me to find parking,” “reporting lines may change”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

-",

-1.19

 指輪

-1.16

succeeded

-1.13

不屑

-1.13

ctory

-1.09

 .......

-1.08

 SCIENCE

-1.05

 gewi

-1.04

"},

-1.04

POSITIVE LOGITS

 sometimes

1.62

 иногда

1.55

 manchmal

1.51

 always

1.34

 Sometimes

1.27

 occasionally

1.23

 still

1.20

 keeps

1.19

 Getting

1.19

 gets

1.18

Activations Density 0.001%