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

</table>

-0.32

Authorized

-0.31

 libres

-0.29

 berlina

-0.28

do

-0.28

↵↵

-0.28

 rowspan

-0.28

Lưu

-0.27

//

-0.27

POSITIVE LOGITS

WSGI

4.38

wsgi

1.44

asgi

1.02

 يتيمه

0.99

 ब्रेकडाउन

0.88

MemoryWarning

0.82

DJANGO

0.81

 betweenstory

0.81

parsedMessage

0.79

 AssemblyTitle

0.77

Activations Density 0.000%

No Known Activations

This feature has no known activations.