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

<eos>

-0.34

ichert

-0.33

łek

-0.32

setData

-0.32

 điện

-0.30

 commission

-0.30

 competition

-0.30

ícola

-0.30

zsef

-0.30

 voulons

-0.29

POSITIVE LOGITS

@"/

7.13

tagHelperRunner

1.59

AndEndTag

1.21

 للاسماء

1.16

 CreateTagHelper

1.13

 ***!

1.09

httphttps

1.07

 autorytatywna

1.04

ValueStyle

1.03

InjectAttribute

1.00

Activations Density 0.000%

No Known Activations

This feature has no known activations.