INDEX

Explanations

Code comments, questions, and statements

The neuron fires on identifiers being pulled out of module imports (e.g. destructured names on the left‐hand side of `require(...)` or `import {…}` statements).

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

 advertencia

-1.23

罴

-1.21

mitsub

-1.15

⎔

-1.14

щихся

-1.14

UREAU

-1.13

鼍

-1.13

薆

-1.11

赑

-1.10

these

-1.07

POSITIVE LOGITS

1.27

1.23

1.17

etches

1.16

1.15

 выбирать

1.13

 говорит

1.08

изнь

1.06

眼前

1.04

стема

1.02

Activations Density 0.002%