INDEX

Explanations

introducing definitions or hypotheses

This neuron activates on LaTeX math‐mode tokens, especially the “\mathcal{…}” constructs and related mathematical symbols.

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

 bluz

-1.06

 chè

-1.06

他们

-1.05

 dieu

-1.05

 feront

-1.04

 HOUSE

-1.02

 isim

-1.02

なのだ

-1.01

explain

-1.01

 zwier

-1.01

POSITIVE LOGITS

 with

1.50

 whose

1.24

If

1.16

With

1.12

for

1.08

There

1.08

anganan

1.06

if

1.05

now

1.03

 want

1.02

Activations Density 0.036%