INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

layoutControl

-0.49

DataMember

-0.39

 iprot

-0.36

 sobie

-0.34

iLocal

-0.30

os

-0.29

uteur

-0.29

 trwa

-0.29

 rambut

-0.29

 Monats

-0.28

POSITIVE LOGITS

DotNetBar

3.58

MENAFN

1.83

Revenir

1.44

parsedMessage

1.21



0.87

tagHelperRunner

0.80

Personensuche

0.79

messageInfo

0.79

exitRule

0.78

styleType

0.77

Activations Density 0.000%

No Known Activations

This feature has no known activations.