INDEX

Explanations

Texas Instruments

The neuron primarily activates on number‐heavy tokens (i.e. digit sequences or numerals).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

nitz

0.51

 spite

0.51

ษัท

0.50

 मनुष्य

0.50

웨어

0.50

 Viele

0.49

 viele

0.49

 many

0.49

 primitive

0.49

 सात

0.49

POSITIVE LOGITS

↵↵↵↵↵↵

0.62

 evaluated

0.59

 rapor

0.58

':'

0.57

↵↵↵↵↵↵↵↵

0.56

 cohom

0.54

$~

0.54

 pptn

0.53

レポート

0.52

르는

0.52

Activations Density 0.026%