INDEX

Explanations

multi-lingual text

This neuron responds to unusual or low‐frequency tokens—e.g. technical jargon, foreign‐language words, and rare Unicode sequences.

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

婶

-1.27

參考文獻

-1.23

Список

-1.17

͐

-1.10

 lägen

-1.10

Опис

-1.08

шому

-1.05

 theres

-1.04

 словом

-1.04

ffè

-1.03

POSITIVE LOGITS

。

1.46

 døde

1.26

 occurs

1.20

 måske

1.14

atique

1.13

raft

1.11

 diferenças

1.10

itats

1.09

 OMITTED

1.09

感じが

1.09

Activations Density 0.002%