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

 nakalista

-0.94

AddTagHelper

-0.66

transQ

-0.66

LabelTagHelper

-0.65

 bParam

-0.59

AndEndTag

-0.59

ThroughAttribute

-0.56

+:+

-0.55

 ब्रेकडाउन

-0.54

JNIEnv

-0.54

POSITIVE LOGITS

MetaObject

4.81

Meta

0.59

 Meta

0.56

 meta

0.48

META

0.46

 META

0.43

meta

0.43

__(/*!

0.42

MetaData

0.39

MOC

0.38

Activations Density 0.000%

No Known Activations

This feature has no known activations.