INDEX

Explanations

months and days of the week

The neuron strongly activates on temporal references—particularly month names and day-of-week words.

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

 idag

-1.33

岁

-1.23

-1.09

 sustancias

-1.09

 Addon

-1.07

 lesquelles

-1.06

本体

-1.06

(“

-1.04

 GOODS

-1.04

Ia

-1.03

POSITIVE LOGITS

It

1.29

 Estaba

1.23

ossa

1.20

 dibuat

1.19

 Según

1.19

 única

1.19

aupun

1.15

tshell

1.13

 Isso

1.11

举动

1.11

Activations Density 0.012%