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

 zugesch

-0.41

or

-0.39

 Byers

-0.35

Otros

-0.34

ูป

-0.34

GroupBox

-0.33

 loses

-0.33

stdarg

-0.32

 उसने

-0.32

tableCell

-0.32

POSITIVE LOGITS

 Illuminate

2.38

Illuminate

1.52

BASEPATH

0.95

InjectAttribute

0.85

mybatisplus

0.83

enderror

0.80

 فريبيس

0.77

endphp

0.75

AccessorTable

0.75

 GenerationType

0.74

Activations Density 0.000%

No Known Activations

This feature has no known activations.