INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

醐

-0.57

AddArg

-0.53

 dads

-0.51

IUrlHelper

-0.48

elry

-0.48

Ã

-0.47

BibitemShut

-0.47

 Protocol

-0.47

/\.

-0.46

AddTagHelper

-0.46

POSITIVE LOGITS

LayoutStyle

0.44

جغرافيا

0.43

RegressionTest

0.39

onomía

0.37

devamını

0.36

">//

0.34

 AssemblyCulture

0.33

 initComponents

0.32

 claramente

0.32

FieldBuilder

0.32

Activations Density 0.000%

No Known Activations

This feature has no known activations.