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

Not in Any Lists

No Comments

Negative Logits

ⓧ

-0.49

锡

-0.49

 nonUne

-0.48

 Sør

-0.47

勘

-0.47

 PeEnEo

-0.47

={()

-0.47

/*

-0.47

Pt

-0.46

 deut

-0.46

POSITIVE LOGITS

extAlignment

0.52

addGap

0.50

WriteLiteral

0.42

awtextra

0.40

addPreferredGap

0.40

expandindo

0.35

 nahilalakip

0.34

voorbeeld

0.34

addContainerGap

0.34

FieldBuilder

0.33

Activations Density 0.000%

No Known Activations

This feature has no known activations.