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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 صوتيه

-0.97

 CreateTagHelper

-0.93

IntoConstraints

-0.91

MLLoader

-0.90

ロウィン

-0.88

OGND

-0.88

ImageContext

-0.88

 Photocase

-0.87

TagMode

-0.85

 kasarigan

-0.84

POSITIVE LOGITS

 Svensson

0.50

0.45

 another

0.44

 ebenfalls

0.43

 also

0.43

0.40

 Herrmann

0.38

 McGowan

0.37

0.36

Activations Density 0.000%

No Known Activations

This feature has no known activations.