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

abbix

-0.38

 waiting

-0.36

SceneManagement

-0.36

Fah

-0.35

 nguy

-0.35

 Hicks

-0.34

el

-0.34

NavigationView

-0.34

ächtig

-0.34

wait

-0.33

POSITIVE LOGITS

DebuggerNonUser

6.91

DebuggerStep

0.91

RenderAtEndOf

0.90

AddTagHelper

0.82

GOTREF

0.80

 defaultstate

0.76

 CreateTagHelper

0.75

ThroughAttribute

0.74

///</

0.71

fjspx

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.