INDEX

Explanations

code snippets and annotations

This neuron detects occurrences of the string “github” (as in GitHub URLs or references).

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

and

-2.14



-1.98

ATIVA

-1.94

noční

-1.87

unduh

-1.83

ksjoner

-1.78

pasti

-1.77

nyt

-1.71

 واتس

-1.71

tieron

-1.66

POSITIVE LOGITS

2.39

1.90

no

1.73

1.69

or

1.69

 mantém

1.62

 fará

1.62

1.60

es

1.59

ok

1.59

Activations Density 0.029%