INDEX

Explanations

technical terms and dates

The neuron turns on for very high‐frequency small tokens—things like apostrophes in contractions, single‐letter words and other common function words (prepositions, “import,” etc.), and punctuation—i.e. it flags “stop‐word” and punctuation tokens.

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

estaba

-0.90

...»

-0.89

 Debre

-0.86

zwei

-0.84

EXIT

-0.83

...).

-0.82

IENCE

-0.82

","");

-0.81

več

-0.81

⌦

-0.79

POSITIVE LOGITS

Buffers

0.82

科普

0.82

 Evangelio

0.81

 equal

0.80

digos

0.79

往上

0.77

ємо

0.76

 andere

0.76

GeneratedCode

0.76

 rating

0.75

Activations Density 0.006%