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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wapV

-0.51

ambique

-0.41

extendable

-0.41

 Motif

-0.40

Referencer

-0.40

moiselle

-0.39

berra

-0.39

appliquent

-0.39

ôle

-0.39

Économie

-0.39

POSITIVE LOGITS

 ActionBar

0.65

 jspb

0.47

ArrowToggle

0.47

rungsseite

0.41

Personendaten

0.40

 ISNI

0.38

̈́

0.36

findpost

0.36

 linkovi

0.35

 незавершена

0.35

Activations Density 0.000%

No Known Activations

This feature has no known activations.