INDEX

Explanations

occurred with

This neuron detects tokens that are part of named-entity spans (especially organization names).

New Auto-Interp

Configuration

Prompts (Dashboard)

238,145 prompts, 512 tokens each

Dataset (Dashboard)

lmsys + oasst1

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

шинство

0.35

 والأ

0.34

፩

0.34

⁤

0.34

يف

0.34

愀

0.34

𝓐

0.33

𒐪

0.33

 İşte

0.33

Neces

0.33

POSITIVE LOGITS

 aforesaid

0.38

 hospitality

0.31

ดังกล่าว

0.30

 spas

0.30

 semen

0.30

 garner

0.29

 privaten

0.29

 huts

0.29

 moneys

0.29

 sauces

0.28

Activations Density 1.073%