INDEX

Explanations

starting at

The neuron fires on standalone numeric hour values in time‐of‐day expressions (e.g. the “1” in “1:11 a.m.,” the “2” in “2:30 pm,” or the “7” in “7:00 pm”).

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

 they

-1.48

𝐃

-1.23

itaji

-1.19

৳

-1.19

簡介

-1.16

 possibilités

-1.16

formazioni

-1.15

</table>

-1.15

咂

-1.13

一会儿

-1.12

POSITIVE LOGITS

and

1.57

 will

1.54

1.44

了她

1.27

了些

1.24

 amongst

1.18

弟弟

1.16

icoot

1.15

了许多

1.12

了很多

1.11

Activations Density 0.001%